Validating fast Fisher Matrix parameter uncertainty estimation for simulations of the Roman Galactic Exoplanet Survey

The Roman Space Telescope will conduct a time-domain survey of the galactic bulge to search for and find over a thousand cold exoplanets using the microlensing technique. To date, estimates of the Roman survey yield using the gulls simulation code have focused only on the question of whether a planet will be detected and have neglected to address whether it will be possible to characterize the planet once detected. We have extended gulls to compute Fisher matrix estimates of the planet parameter uncertainties obtainable with Roman observations of individual microlensing events. Here we focus on validating the Fisher matrix calculations with uncertainty estimates derived from Markov Chain Monte Carlo (MCMC) posterior probability distributions. We will present results for the accuracy of fisher matrix parameter estimates for a series of more challenging microlensing models: single lens events, free-floating planet events, and ultimately bound planet events. For the latter two cases, we will estimate Roman’s ability to measure planet masses in simulated events. Ultimately, validated, fast Fisher matrix uncertainty estimates will allow us to optimize the Roman Galactic Exoplanet Survey for characterizable planets rather than simply detectable planets.