3. Data analysis
3.1. DEM analysis with CDS
The line intensities reported in Table 1 have been used to determine the coronal hole DEM , defined as
Element abundances have been taken from Feldman et al. 1992 (coronal values). Ion fractions come from Arnaud & Raymond 1992 (Fe ions) and Arnaud & Rothenflug 1985 (other ions). To determine the DEM of the emitting plasma we used the iterative method described by Landi & Landini 1997 to which we refer for details.
The DEM curve is usually defined in a temperature range including plasma from chromospheric to coronal conditions; above the coronal temperature, where no more lines are detected by the instruments, the DEM function is very often extrapolated to a very low value at a very high temperature (). This causes the DEM to have a sharp maximum around the coronal temperature. Landi & Landini 1998 have criticized this assumption, since a decreasing DEM at high temperatures implies an increasing temperature gradient , which would produce a physically unacceptable high temperature. It would also pose problems to the energy balance equations, because the temperature gradient is linked to the heat conductive flux. Consequently, Landi & Landini 1998 proposed to truncate the DEM at a maximum temperature beyond which the DEM is not defined; in this way does not increase when the temperature approaches .
In this paper we will use both methods to determine the DEM from the observed EUV line intensities in the coronal hole. These DEM curves will be then tested by computing the radio brightness temperature and comparing it with the observations.
The DEM fits are displayed in Fig. 3 together with the data points. It must be pointed out that, in the data points plotted in the figure, the Fe, Mg and Si abundances from Feldman 1992 have been lowered by a factors ranging from 3.4 to 4.0, getting a good agreement with the Grevesse & Anders 1991 photospheric values. This indicates that the First Ionization Potential effect (FIP effect - see e.g. Haisch et al. 1996) cannot be applied to this dataset, in agreement with the findings of Del Zanna & Bromage 1999.
3.2. Radio brightness temperature
is the optical depth of the level where and . At very short wavelengths, where the refractive index in the TR and corona, can be directly derived from the EUV lines intensity through the DEM without any further assumption and the integral in Eq. 3 is performed up to
At the frequencies considered in this paper this approximation does not hold, as the critical density is located in the high TR or in the corona.
In order to take into account the refractive index n in Eq. 4, we must know the electron density profile in the upper part of the solar atmosphere. This has been done in the two following ways:
The lower temperature in the above integral has been arbitrarily set K: this value does not affect the calculations since all considered radio frequencies have their critical level at temperature and therefore the integral of the transfer equation stops at higher temperature.
Pressure values lower than would give at .
© European Southern Observatory (ESO) 1999
Online publication: July 16, 1999