Spectral fitting can get very complicated very quickly (e.g., see Figure 17). There are typically six or seven spectral components that should be fit to the data, even before the spectrum of the object of interest (if that is not the cosmic background) is added. These components are split between instrumental and cosmic origins. Some of the components can vary with time, observation direction on the sky, position on the detector, location of the satellite in its orbit, and location of the Earth in its orbit.
As alluded to above, even if your primary interest is imaging, you will still need to do some spectral fitting to determine the contributions from the residual soft proton flares and solar wind charge exchange (SWCX). If you have only a single observation, disentangling the solar wind charge exchange emission from your source spectrum is likely to be impossible. If you have multiple overlapping observations, determining the SWCX and residual soft proton flare (SPF) emission becomes more certain. In the case of imaging analysis, the fitting of the source spectrum may not need to be as physically meaningful as it would be for a spectral analysis, it just needs to reproduce the underlying source spectrum well enough to get a good measure of the residual soft proton flare and SWCX emission
The following sections describe the peculiarities of spectral fitting for ESAS extracted spectra. Those concentrating on imaging analysis may not need to read more than the first several parts. For imaging analysis, fitting in the “traditional” manner with subtracted background will probably be sufficient. More rigorous spectral analysis will probably require simultaneous fitting of the source and the background. ESAS is not yet creating the background spectra in a way that such fitting can be done with the C-statistic, but that is being developed.