A New Method for Doppler Imaging of Starspots: Sequential Linear Least Squares Estimation

Previous abstract Next abstract

Session 45 -- Early Type Stars
Display presentation, Wednesday, 9:20-6:30, Heller Lounge Room

[45.09] A New Method for Doppler Imaging of Starspots: Sequential Linear Least Squares Estimation

S.L. Allen, S.S. Vogt (UCO/Lick, UCSC)

We describe a new technique for deriving Doppler images of stars. It uses sequential linear least squares estimation to reconstruct the image by inverting a set of high-resolution spectral line profiles. An attractive feature of this method is that the entire dataset need not be inverted simultaneously. This allows the inclusion of arbitrarily large observation vectors such as are obtained with echelle spectrographs. In contrast to the maximum entropy method, least squares also provides formal error estimates which indicate how much confidence should be attributed to various features in the resulting starspot maps.

We also discuss an improved geometrical model based upon the binary star work of A.P. Linnell. This model incorporates the effects of the Roche distortion of these spotted stars caused by their close binary companions. Well-studied spotted stars such as UX Ari are significantly tidally distorted (almost to the point of Roche-lobe overflow). Most previous studies have omitted the effects of non-spherical shape, variation of surface gravity, associated gravity darkening and mutual irradiation.

This work has been partially supported by NSF Grant AST-9115376.

Wednesday program listing