The principal tools for analyzing the splitting data are being incorporated into an Inversions WorkBench. This is currently being developed in IDL, but other environments might also be used in the future. This provides a widget-driven, graphical interface to a number of inversion programs. The principal inversion methods are all linear, and utilize least-squares fitting (possibly with variants such as truncating or rolling off a singular value decomposition), optimally localized averaging (or other designs of linear averaging kernels), or 1x1 decomposition. We shall use full 2-D inversions based on regularized least squares and 1x1 decomposition (some of these are already incorporated into the WorkBench), 2-D asymptotic inversions, optimally localized averages and other designs of averaging kernels to look for features such as gradients and shear zones. A further class of inversion methods that will be used to probe the two-dimensional variation of the rotation are so-called 1.5-D or semi-parametrized methods: these are currently being incorporated into the WorkBench. It is also useful to make cuts through the solar interior using one-dimensional inversions, for example in the equatorial plane. These have the attraction that they are fast and utilize well-developed methods: least-squares with regularization or filtering of singular values, optimally localized averages and other designs of averaging kernels in one dimension, asymptotic inversions.
In addition, but not presently envisaged to take place within the WorkBench, forward modelling with parametrized rotation models will also be used to investigate the solar rotation.
This page last reviewed and revised 14 Aug 1995
Please address comments and questions to the Team Coordinator
Return to SOI Main Page
SOI Team Science page