The Velocity Differential Emission Measure of Solar Flare Soft X-Ray spectra

fl190.newton01
Posted:  30-Sep-94
Updated: 05-Nov-94, 25-Apr-95, 15-Nov-95, 03-Feb-96, 1-Sep-96
Events specified: Flares including 9-Nov-91 03:20 and 5-Jan-92 13:15

E. Newton and A. G. Emslie (University of Alabama in Huntsville), and J. T. Mariska (NRL)

Chromospheric evaporation is thought to occur as part of the solar atmosphere's response to the deposition of energy in the upper chromosphere during a flare. Two-component models are generally used to test for chromospheric evaporation evidence in soft X-ray spectral line profiles. We have determined, however, that such models are inconsistent with observed spectral lines and that oftentimes the fitting of two components can lead to erroneous conclusions regarding the presence of chromospheric evaporation. Recognizing the shortcomings of two component models, we have derived a new quantity called the velocity differential emission measure (VDEM). The VDEM is directly related to the flaring plasma's properties and measures the amount of emission from a volume of plasma moving with a given velocity. It can either be computed from its definition, or recovered by inverting spectral line profiles. The VDEM can provide direct evidence for chromospheric evaporation. We plan to study VDEM's computed for a number of flares observed with the BCS on Yohkoh, as well as spectra obatined from numerical simulations, with the goal of understanding better the dynamical processes taking place early in a solar flare. Initially, we are working with flare data that has already been analyzed by others, including the 1991 Nov 9 and 1992 Jan 5 flares. As we become confident in our inversion process, additional BCS flares will be added. This work will form the core of E. Newton's PhD thesis at the University of Alabama at Huntsville.

Update 1-Sep-96

I would like to submit to you the abstract (attached) of the paper I will be submitting to ApJ at the end of August 1996.

INVESTIGATING ``PRECURSOR FLOWS" IN SOLAR FLARES

Despite the considerable theoretical and predictive success of the electron-heated model of solar flares, observations of low levels of soft X-ray emission {\it prior to} the detection of hard X-rays in most flares present a problem for the paradigm, since it suggests some sort of ``pre-heating" of the chromospheric plasma before electron acceleration. Particularly troublesome are assertions that plasma {\it flows} have been observed prior to hard X-ray emission, because they imply significant and rapid energy input by some agent without leaving a hard X-ray signature. Given this challenge to an otherwise successful model, it is imperative to investigate more comprehensively the existence of ``precursor flows" manifested in the asymmetry of X-ray spectral lines. In this paper, we examine a sample of 50 disk-center flares observed by the {\it Yohkoh} BCS between 1991--1993 and determine with the Velocity Differential Emission Measure (VDEM) the extent to which ``precursor flows" occur. We explore specifically the possibility that these ``precursor flows" are only {\it apparent}, arising from the existence of more than one emission region on the solar disk. The spatial separation between these regions causes the emission to be registered by the detector in a wavelength bin which is shifted relative to the line's nominal location, giving rise to line asymmetries. In our sample, we find that only five flares (1991 November 9, 1992 January 30, 1992 February 29, 1992 April 7, and 1993 April 22) manifest ``precursor flows," all of which can be explained by the presence of other active regions. We conclude, therefore, that ``precursor flows" do not exist in our sample and that the supposed challenge to the electron-heated paradigm is not as significant as previously asserted.

Update 03-Feb-96

The dissertation containing my research to date on the velocity differential emission measure (VDEM) has been completed. Please refer to the thesis section of the bulletin board for a summary.

Investigations involving VDEM, however, continue. Most recently, I am currently drafting a paper for submission to the Astrophysical Journal concerning observations of ``precursor" flows of soft X-ray emitting plasma *prior* to the detection of hard X-rays, which is quite problematic for the canonical thick-target electron beam model of flare heating. In an analysis of 46 disk-center flares, only 5 showed possible evidence for such ``precursor" flows. Closer examination of these flares, however, reveal that the supposed plasma flows are only apparent, rather than actual, Doppler motion, as manifested in VDEM for these flares. The apparent motion evidently arises from the presence of emission from an X-ray source located south of the region identified as the primary flare site. Such a spatial offset (north/south) along the BCS's crystal dispersion axis results in an apparent motion of 1.1 km s^-1 per arc sec. Examination of SXT's observing logs confirm the existence of these additional flare regions.

Update 15-Nov-95

Abstract of paper accepted for publication in the Astrophysical Journal, March 10, 1996 issue:

"Testing the Impulsiveness of Solar Flare Heating through Analysis of Dynamic Atmospheric Response"

One crucial test of a solar flare energy transport model is its ability to reproduce the characteristics of the atmospheric motions inferred from soft X-ray line spectra. Using a recently developed diagnostic, the velocity differential emission measure (VDEM), we can obtain from observations a physical measure of the amount of soft X-ray emitting plasma flowing at each velocity v, and hence the total momentum of the upflowing plasma, without approximation or parametric fitting. We have correlated solar hard X-ray emission profiles observed by the Yohkoh Hard X-Ray Telescope with the mass and momentum histories inferred from soft X-ray line profiles observed by the Yohkoh Bragg Crystal Spectrometers. For suitably impulsive hard X-ray emission, an analysis of the hydrodynamic equations predicts a proportionality between the hard X-ray intensity and the second time derivative of the soft X-ray emitting plasma's momentum. This relationship is borne out by an analysis of 18 disk-center impulsive flares of varying durations, thereby lending support to the hypothesis that a prompt energy deposition mechanism, such as an energetic electron flux, is indeed responsible for the soft X-ray response observed in the rise phase of sufficiently impulsive solar flares.

Further analysis employing VDEM is underway. In particular, preliminary results indicate that it is possible to correlate the information on the line-of-sight velocity distribution (represented by VDEM) with SXT images (which convey information on transverse velocities) to construct a three-dimensional picture of plasma flows. In addition, the combination of the two datasets permits ambiguities between proper versus apparent motion to be resolved. A manuscript is currently being prepared on this subject.

Update 25-Apr-95

We have continued to refine our inversion process to deconvolve observed Ca XIX line profiles and recover their Velocity Differential Emission Measures (VDEMs). In particular, we have incorporated the effect of a known satellite line feature (d13) into our analysis so that VDEMs reflect solely emission from the Ca XIX resonance transition. We have also implemented a method of choosing the optimal smoothing parameter for the inversion by minimizing the Generalized Cross-- Validation function. The smoothing parameter is determined for each individual soft X--ray spectrum, as each dataset has different noise levels associated with it. We are also able to compute the uncertainties on the recovered VDEMs by utilizing the covariance matrix of the solutions.

Approximately 50 Sun--centered solar flares, observed by Yohkoh, have been selected for further analysis using our technique. Movies of VDEMs' evolution over the course of a flare have been made. In addition, the timing correlations between hard X--ray bursts and VDEM features are being explored, as well as mass and momentum continuity in the solar atmosphere's response to flare energy input. We have determined, through an analysis of the hydrodynamic equations governing the response of the solar chromospheric plasma to localized electron beam heating, that the soft X-ray plasma's momentum and its temporal derivatives should be related to the strength of the corresponding hard X--ray burst. We are currently exploring this theoretical relationship and searching for evidence in data observed by the BCS and HXT.

A manuscript reporting our data analysis is currently in preparation. Our earlier work, which established the theoretical basis and viability of our inversion technique, will appear in the July 10, 1995 edition of The Astrophysical Journal.

Update 5-Nov-94

ABSTRACT OF PAPER SUBMITTED TO THE ASTROPHYSICAL JOURNAL

"The Velocity Differential Emission Measure: Diagnostic of Bulk Plasma Motion in Solar Flares"

Elizabeth K. Newton, A. Gordon Emslie, John T. Mariska

Mass motions are a ubiquitous product of solar flare energy release. A better understanding of the flare plasma's distribution--how much is moving and how fast--permits insight into the mechanisms of energy transport (and release) which lead to those motions. Observationally, mass motions during flares are often manifested in the shape and location of soft X-ray emission lines. Observed lines generally exhibit a width greater than the thermal Doppler width and a blue-wing asymmetry which has been cited as evidence for plasma motions along the line-of-sight. Past efforts to characterize this excess width and asymmetry have primarily involved the parametric fitting of a double-gaussian form. In this paper we show, however, that simple two-component models are inconsistent with the observed evolution of spectral lines and hence serve as a poor diagnostic of plasma motions.

We therefore generalize the synthesis of line profiles to the case of a continuum of gaussian components, by introducing a quantity

which we term the velocity differential emission measure (VDEM).  The
VDEM measures the distribution of emission from a volume of plasma as
a function of its line-of-sight velocity.  It can either be computed from
theoretical model atmospheres, or recovered from observed line profiles

Work on this topic continues. We are currently developing time series of VDEMs for as many flares as possible. We will be looking for common trends in the VDEMs, as well as examining hard X-ray data for correlations between its timing and changes in VDEM. The measuring of momentum balance with VDEM is also being explored.