Worksheet 6 - Fitting Gaussians to the data #
Generally the EIS emission lines have a Gaussian shape and, by fitting a Gaussian function to the spectrum, one can derive line intensity, line width and velocity maps. The high sensitivity of EIS means that for many lines it is possible to perform good quality fits at each pixel in the image. The routine eis_auto_fit in Solarsoft takes the output from eis_getwindata (Worksheet 5) and fits a single Gaussian to each pixel in the image. It is called as
IDL> eis_auto_fit, wd, fit, refwvl=195.12
refwvl is used to specify a reference wavelength for the emission line, so that velocities can be calculated. Reference wavelengths for lines can be obtained from the CHIANTI database, e.g.,
IDL> which_line,’fe_12’,195
If a wavelength window contains more than one emission line, then it is possible to use a reduced wavelength range with the wvlpix= keyword. E.g., wvlpix=194.9,195.4 to use only pixels between 194.9 and 195.4 Å.
To see what the output structure contains, do:
IDL> help,fit,/str
the tags int, cen, vel and wid correspond to intensity (units: erg cm-2 s-1 sr-1), centroid (units: Å), velocity (units: km s-1) and Gaussian width (units: Å). To get the FWHM, multiply the Gaussian width by 2.35. For each of the tags, there is an associated 1σ error array.
The results of the fitting process can be checked using the widget-based routine eis_fit_viewer:
IDL> eis_fit_viewer, wd, fit
The top 3 graphic windows in the GUI show line intensity, line velocity and line width (FWHM). Note the vertical bands in the velocity window – this is due to the orbital variation of the lines on the detector (see Worksheet 7a).
With the ‘Zoom’ button depressed, the mouse can be used to draw a ‘rubber-band’ box to zoom into the images. With the ‘Pixel’ button depressed, the line profile from the selected pixel will be drawn in the bottom-left graphic window.
The bottom-right window shows a histogram of the selected images pixels for either intensity, velocity or line width.
Exercise #
- Some horizontal stripes can be seen in the velocity and line width maps. What could be causing these?
- Use eis_fit_viewer to find the approximate variation in line centroid position caused by the orbital motion of Hinode.
