We have presented a model of the GOLF system which allows the calculation of the instrument response to specified spatial velocity functions and given sensitivity functions for use with the evaluation of magnetic field effects. This model requires quantitative values for a number of instrumental parameters including the scattering strengths and the instrument parasite light. The values used are derived from measurements during preflight tests or from an analysis of the data returned from the spacecraft. The comparison shown in the previous section between the model as a function of and the values deduced directly from the GOLF observations provides a critical validation of the chosen parameters.
The approach we have adopted in previously published papers (Henney et al. 1999, Bertello et al. 2000a, 2000b) uses the signal cross-correlation method instead of the signal subtraction method. The cross-correlation approach is much less demanding on the modelling theory than would be the subtraction method. We describe here some of the extensions which would be desireable in an application of the signal subtraction method should future instrument systems provide adequately low-noise data.
The velocity and magnetic field effects discussed in this paper must be joined by effects due to temperature and intensity fluctuations which are part of the oscillatory and convective processes. For the coherent oscillations from the GOLF instrument, this question has been studied by Pallé et al. (1999) and by Renaud et al. (1999). These studies show that for the higher frequency coherent oscillations, the signal due to the thermal or excitation effects in the solar atmosphere produces only small alterations in the phase relationships that one would get from an instrument which measures only velocity. However, the convective and magnetic phenomena which can also contribute to the GOLF single wing signal through the intensity may have a completely different nature. As part of the synoptic program at the 150-foot solar tower on Mt. Wilson, daily intensity images are made available on the web page at http://www.astro.ucla.edu/~obs/intro.html . These are line bisector images for both and Na D1. The bisector intensity as shown by these images is clearly altered by the presence of both weak and strong magnetic fields. In addition, outside of the magnetized areas, the intensity shows variations on a spatial scale comparable to the supergranulation or chromospheric network. These features appear to be convective in origin and are probably a component of the low frequency noise which interferes with the search for coherent solar oscillations.
The Mt. Wilson 150-foot solar tower system now includes regular observations of the nm line in a manner which permits the simulation of the MDI signal. A comparison between images of the MDI line depth parameter and the D1 bisector intensity suggests that the MDI line depth can be used to derive the convective intensity variations. Efforts to quantify this possibility are currently in progress. The MDI instrument also provides a continuum flux index which can be used to identify sunspot regions and treat their contribution in a different but appropriate manner. Sensitivity functions like those of Eqs. (35) and (36) can be written using a multiple parameters like the MDI line depth and continuum flux observables instead of . As long as appropriate coefficients are known to convert these parameters into a resulting line bisector intensity change, the equations can be applied immediately to the simulation. We know from Ulrich et al. (1993) that the primary effect on a resonance scattering cell signal can be modelled. It will require a detailed application to the GOLF data to determine the frequency range of validity of this model.
The inclusion of additional information derived from the available
data bases will make it possible to carry out a simulation of the GOLF
signal which takes into account a variety of non-coherent solar
phenomena. If the simulation is successful, it may be possible to
enhance the ratio of coherent oscillation signal relative to the
non-coherent contributions and improve the detectability of weak solar
oscillations. The MDI instrument has available data products which can
be applied to this project: First are two velocity sequences - the
medium velocities which have good
spatial resolution and are at the full temporal cadence of MDI but
omit data from the solar limb regions and contain some temporal gaps
due to the telemetry mode and other programatic factors, and the
VIRGO/LOI mask sequence which has lower spatial resolution but
includes the limb zone and has better temporal coverage. Second, are
two non-velocity measures of the solar surface - the line depth
parameter and the continuum flux parameter. Both these are available
only in the form of gaussian smoothed temporal averages of
128128 arrays. The temporal sampling
is once every 12 minutes and the temporal smoothing is done over 23
images using a gaussian weight having a standard deviation of 3.4
minutes. One bad image out of the 23 will spoil the telemetered image
so this series includes a high number of missing frames. Nonetheless
these sequences permit removal of some of the variations detected by
GOLF which are produced by identifiable solar processes that are not
© European Southern Observatory (ESO) 2000
Online publication: January 29, 2001