Systematic uncertainties in photoionization modeling

Atomic data such as transition energies and rates, used by collisional and photoionization plasma codes, are generally subject to systematic uncertainties. However in many cases it is challenging to characterize these uncertainties reliably. And it is difficult to propagate these uncertainties through the plasma codes to understand their effects on the astrophysical results and their contribution to the total error budget. Here we present an alternative approach to quantifying systematic model uncertainties that is based on comparison of statistically high-quality observational data with those plasma models. That is, we examine the hypothesis that the residual discrepancies from fitting to high statistical quality spectra are entirely due to systematic uncertainties in the atomic data. We use archival Chandra/HETG data of the Seyfert I galaxy NGC 3783 and fit the xstar/warmabs photoionization model. We examine systematically the data-model discrepancies and how they can provide feedback on the quality of the underlying atomic data. We address the accuracy of the charge state distribution and individual line optical depths and how to account for these during fitting in terms of a model uncertainty.