CossamDoppler

ZDM papers abound in claims for often spectacular abundance patches in magnetic ApBp stars, usually in stark contradiction with what diffusion theory predicts; the blame for the discrepancies has always been put squarely on theory. More recently however, many of these claims have been challenged and imputed to various shortcomings of the ZDM approach used.

Few tests have been published concerning the capabilities of the most popular ZDM code, restricted to the recovery of perfectly symmetric, vertically constant (unstratified) abundance spots. Lately, a ring-like abundance structure following the magnetic equator of a centred dipole has also been considered, but still with unstratified abundances. On the other hand, alternative tests with CossamDoppler revealed difficulties when it comes to the recovery of warped rings in non-axisymmetric magnetic geometries, but also of "spots" seriously deviating from perfect symmetry, and in particular of stratified abundances that depend on magnetic field angle and strength.

Surely, for the interested reader of the relevant ZDM papers it must be difficult to assess the meaning and credibility of these opposing claims. Whereas on the one hand the relevance for real stars of those highly idealised spots and/or rings (always with stellar parameters "optimal for DI" in the words of the authors) may be questioned, sceptics on the other hand are entitled to suspect a biased emphasis on a handful of extreme cases.

The most popular magnetic Doppler mapping code used in recent papers on ApBp stars is not freeware. It is known that learning how to work with this code is non-trivial and time consuming, maybe equivalent to about half of ones PhD thesis or, perhaps, several months of work for an experienced spectroscopist. You are expected to run the Fortran code on MPI-enabled supercomputer clusters. Even though access to this particular ZDM code is possible in the context of collaborative projects (where all stages from problem formulation to eventual publications have to be negotiated) it is hard to imagine that this could result open-ended as regards the outcome and allow an independent assessment of strengths and weaknesses of present-day Zeeman Doppler algorithms.

CossamDoppler remedies this situation. It is an easy-to-use code with a reasonably detailed manual, making it possible for everyone to make up her/his own mind on ZDM. There is no need for MPI or OpenMPI; the free GNAT Ada compiler provides all the required parallel constructs (albeit restricted to SMP architectures). You don't have to specify and negotiate tedious details of a joint project, you are indeed completely free to carry out tests with any vertical and horizontal abundance distributions of your liking, thanks to Stokes profiles established with the help of CossamMulti. You can play with multiple simple spots, asymmetric spots with complex inner structure, ring-like local features, warped rings following the magnetic equator of a tilted eccentric dipole ... It is possible to have a close look at what happens when the magnetic field is neglected altogether or just moderately in error. You can get a feeling of the spatial resolution that can be achieved and how far the positions of various features can be correctly recovered.

CossamDoppler comes with a whole set of auxiliary plotting programs for the visualisation of the results. Don't hesitate to contact me for any suggestions as to improvements or for bug reports.

CossamDoppler has evolved over the years 2013-2017. For the history of Cossam see the Cossam page.

The present version dates from February 2, 2017.

Copyright 2017 Martin J. Stift

object-oriented, parallel, and written in Ada