Spectra at each spatial pixel in a raster can be fit with Gaussians using the eis_auto_fit IDL routine developed for the Hinode/EIS mission (you must have the EIS branch in your Solarsoft distribution). The key step is to format a SPICE wavelength window into a “windata” structure using the routine spice_getwindata. An example call is:
IDL> wd=spice_getwindata(file) ; choose a wavelength window
IDL> eis_fit_template,wd,template ; create a template for the fit
IDL> eis_auto_fit,wd,fit,template=template
IDL> eis_fit_viewer,wd,fit
The routine eis_fit_viewer is a GUI-based routine for browsing the fit results. More details on the eis_auto_fit suite of routines are available in EIS Software Note 15.
The EIS software enable a number of operations to be performed on windata structures, as described in EIS Software Note 21. For example, eis_bin_windata allows spatial binning of the rasters.
If you have problems running the EIS software on SPICE data, please contact Dr. Peter Young.
In this scenario, a spatial region is selected in the raster, perhaps a bright point or a loop, and the spectra are averaged over this region to produce a single, 1D spectrum. Gaussians are then fit to this spectrum. The procedure is as follows:
IDL> map=spice_make_image(file,770.4) ; for Ne VIII 770.4
IDL> mask=pixel_mask_gui(map) ; GUI routine for selecting pixel mask
IDL> spec=spice_mask_spectrum(file,mask) ; create averaged spectrum
IDL> spec_gauss_spice, spec ; GUI routine for fitting Gaussians
These routines are modified from routines originally written for Hinode/EIS. Further details can be found in EIS Software Note 15.
The fitting templates created for eis_auto_fit can be useful for fitting mask spectra as well. First create the template for the window you are interested in:
IDL> wd=spice_getwindata(file)
IDL> eis_fit_template,wd,template
Now create the mask spectrum using the procedure described above. You can then apply the template to the mask spectrum by doing:
IDL> fit=spice_mask_auto_fit(wd,spec,template,chi2=chi2)
The output structure fit contains the fit results. The reduced chi-square value for the fit is returned in the optional output. To overplot the fit on the spectrum, do:
IDL> p=spice_plot_mask_auto_fit(spec,fit)