Presentation #430.02 in the session Exoplanet Direct Imaging.
Future missions like Roman, HabEx and LUVOIR will directly image exoplanets in reflected light. While current near infrared direct imaging searches are only sensitive to young, self-luminous planets whose brightness is independent of their orbital phase, reflected light direct imaging will observe planets that change in brightness over the course of an orbit due to phase variations. One of the first objectives will be determining the planets orbit via astrometry, the projected position of the planet with respect to its host star in the sky plane. We show that the changing brightness of a planet throughout its orbit can significantly improve the accuracy and precision of numerical orbital retrieval with 2 and 3 direct images. This would speed up the classification of exoplanets and improve the efficiency of subsequent spectroscopic characterization of their atmosphere and surface. We develop a forward model to generate synthetic observations of the three observables produced by a direct image at a given epoch: the two dimensional position of the planet with respect to its host star on the sky plane, and the planet/star flux ratio. Synthetic data are fitted with Keplerian orbits and Lambertian and non-Lambertian phase variations to retrieve orbital parameters, the size of the planet, and its geometric albedo. Our analysis indicates that for astrometric uncertainties of 0.01 AU in projected separation and uncertainties at 10-12 in the flux ratio, using photometry in orbit retrieval improves the precision of semi-major axis by 40% for 2 epochs and 73% for 3 epochs if we use a Lambertian phase curve. With a non-Lambertian phase curve, this improvement is 16% for 2 epochs and 50% for 3 epochs.