Authors: Abhinav G. Iyer (The University of Sydney), Michael S. Wheatland (The University of Sydney), Yang Liu (Stanford University)

Nonlinear Force-free Field (NLFFF) models are often used to investigate the structure of coronal magnetic fields in solar active regions. However, there are limited ways to validate the models. In this work, we use Gaussian separation, recently applied to solar vector magnetogram data, to investigate the accuracy of NLFFF models constructed with two methods : optimisation and the current-field iteration (CFIT) implementation of the Grad-Rubin method. Gaussian separation allows a partition of the total photospheric vector magnetic field into three components, associated with current sources below, above and passing through the photosphere, respectively. Comparing the components of the photospheric magnetic field due to coronal currents in the NLFFF models with those in the original vector magnetogram boundary data, used as a basis for the model, provides a check on whether the NLFFF models accurately represent the coronal currents. We consider NLFFF models constructed for the active regions AR 11429 and AR 12673. We find that the photospheric signatures of the coronal currents derived from the models, for both active regions, are qualitatively similar, but have distinct differences, to those obtained from the vector magnetogram boundary data. These differences are due in part to the changes made in the models to the vector magnetic boundary data when solving the NLFFF equations. The photospheric signatures for both the models and the vector magnetogram data indicate coronal currents flowing above and parallel to the central polarity inversion line (PIL) in both active regions, consistent with recent studies. We propose Gaussian separation as a useful tool to validate coronal magnetic field models, in addition to existing methods. We also discuss how information from Gaussian separation might be directly incorporated into the construction of NLFFF models.