Closure phases are critical in astronomical interferometry. However, their uncertainties are difficult to compute numerically. We provide methods to efficiently compute interferometric closure phase distributions that are valid in the low signal-to-noise ratio regime. We do this by first showing that the von Mises distribution is a good approximation to the full visibility phase distribution, and then performing a triple convolution to obtain an approximation of the closure phase distribution. In order to further accelerate numerical computations of closure phase distributions, we provide techniques to efficiently compute higher order modified Bessel functions of the first kind necessary in the computation of von Mises distributions. The resulting approximations perform much better than the Normal distribution for a wide range of signal-to-noise ratios and, being fully analytic, allow for fast computations that enable them to be incorporated in statistical algorithms.
|Original language||English (US)|
|State||Published - Sep 10 2019|
ASJC Scopus subject areas