Obituary for Dr. Michael R. Rosa (1953 – 2022)

Contributed by Jeremy Walsh, Dietrich Baade, Paul Bristow & Florian Kerber, European Southern Observatory


Dr. Michael Rosa obtained his Ph.D. from the University of Heidelberg in 1981, working at the Landessternwarte Königstuhl on observational analysis of the giant H II region NGC 604 in M33 with Immo Appenzeller as advisor. He joined the European Southern Observatory (ESO) in Garching as a Fellow in 1982 and began as one of the first staff of the Space Telescope - European Coordinating Facility (ST-ECF) at the end of 1984. The ST-ECF, sited at ESO, was initiated through an agreement between the European Space Agency (ESA) and ESO as part of the 15% ESA contribution to the Hubble Space Telescope mission, which included the Faint Object Camera and staff support at Space Telescope Science Institute (STScI), Baltimore. The ST-ECF was founded as a mix of astronomers and software specialists dedicated to HST instrument support, reduction and analysis software, and archive development. Michael joined on an ESA position and remained in the ST-ECF until its closure in 2010. Subsequently he worked at the ESA European Space Astronomy Centre (ESAC) near Madrid on the HST archive and, later, the ESA Sky interface for visualisation of astronomical data, before retirement in 2016.

Michael well embodied the symbiotic nature of the ST-ECF as an independent branch of ESA and ESO, contributing to both organizations. He was very interested in astronomical instrumentation - building an optical and near-infrared photoelectric photometer for his Master’s thesis and, at ESO, investigating the non-linearity of the Image Dissector Scanner (IDS). He also contributed tasks to the ESO MIDAS data reduction system. His first ST-ECF project was in developing the ST Model, a simulator for the 1st generation HST imaging and spectral instruments. This was far more than an exposure time calculator, rather a simulator capable of producing images and spectra using the set of components of the instruments and a database of spectra. When HST was launched in 1990, Michael's prime responsibility was support for the Faint Object Spectrograph (FOS) - a small aperture UV-red spectrograph with a range of spectroscopic resolutions, also high time resolution and spectropolarimetry capabilities. Michael quickly noted that the red part of the spectrum was compromised by extraneous light, not predicted by the instrument simulations. He sought to explain this 'red leak' and, in keeping with the unflinching faith in HST that was shown when spherical aberration was discovered, mitigate its effect on spectra. In retrospect this, at the time, ambitious goal marked much of Michael's later work in spectroscopic instrumentation.

After diagnosing the FOS scattered light problem, Michael produced software to correct its effects on observations. It was at this time that he visualised the power of instrument modelling - using optics (ray tracing) and physics of dispersing elements and detectors to simulate the recorded spectrum for a given object. He sought to take this further and to use the physical models to drive the data reduction software - ultimately to extend to forward modelling, where the only known is the parameterization of the instrument through as complete a model as possible. Then the spectra of the feasible set of targets entering the instrument could be seen as a set of unknowns to be solved-for based on the detected data. This very ambitious goal has not yet been realised in practice but the ever increasing speed of modern computers, advanced structural, thermal and optical modelling capabilities and the application of Bayesian methodology is bringing such a vision closer to fulfilment.

Michael continued to work on FOS calibration, leading to better calibrated data entering the HST post-operation archive, and extending to wavelength calibration of the 2nd generation Space Telescope Imaging Spectrograph (STIS). He formed a group (ST-ECF Instrument Physical Modelling Group) to apply the methodology more generally, also to ESO Very Large Telescope (VLT) instruments. While the instrument physical modelling had begun on existing instruments, the experience developed could be brought into the design process for new instruments to ensure high fidelity data, also allowing manufacturing, assembly and alignment problems to be revealed. This approach has been successfully applied to the VLT instruments CRIRES (CRyogenic Infra-Red Echelle Spectrograph) and X-shooter. In the case of X-shooter the application of a physical model permitted wavelength calibrations to be adapted to changing thermal conditions and gravity vector in a way that is not possible for conventional empirical wavelength solutions.

Given that the FOS scattered light found in the red was so dependent on the target's blue colour, Michael turned his interest to the calibration reference data required to diagnose such problems observationally, such as standard star spectra and laboratory wavelength data. For STIS, the accuracy of the wavelength calibration depends critically on matching the spectrum of the on-board comparison lamp. The existing spectrum of the Pt/Cr-Ne lamp in STIS was incompletely characterised and Michael set about improving this situation by instigating dedicated laboratory measurements at the US National Institute of Standards (NIST) in Maryland, resulting in more accurate wavelengths. This and the modelling work was distinguished by a NASA Group Achievement Award presented to the ST-ECF group together with three NIST colleagues in 2006. His group continued to work on wavelength calibration lamps by contributing to better determinations for hollow cathode lamps, which are used in the near-infrared, important for the high wavelength dispersion of CRIRES. Towards the end of his career, Michael’s experience and attention to detail led to him being chosen to chair several instrumentation reviews at ESO. He brought a disciplined and meticulous approach that contributed to the evolution of the review process at ESO, necessary for the increasingly complex projects of the 21st century.

Michael continued his interest in extra-galactic H II regions, analysing data from IUE, HST and ESO telescopes, publishing extensively on the Magellanic Clouds. He also contributed to study of a range of other targets, such as SN 1987A (discovery of light echoes), SN rates and progenitors, η Carinae, Wolf Rayet nebulae, Sakurai's Object (V4334 Sagittarii) and more generally on abundances in HII regions. One problem to which he made a seminal contribution was to photoionization modelling of nebulae. Until about 25 years ago photoionization models solved the heating and cooling processes of a gaseous nebula considered only in 1D (or pseudo 3D). Michael realised that models needed to be in true 3D and took on a student (Susanne Och) to study the practicality of this idea; he got Leon Lucy interested in the problem and the solution using energy packets and Monte Carlo was established. This approach has been used in the photoionization codes MOCASSIN and Monte Carlo MAPPINGS. With these developments Michael could then witness the advent of more realistic modelling of nebulae, considering non-spherical morphologies and the presence of inhomogeneities.

Given Michael’s zeal to get the very best from astronomical instrumentation, his admiration for the work of Tycho Brahe was not surprising. Together with his Latin teacher from school, he began translating Brahe’s “Astronomiae Instauratae Progymnasmatainto English (“Introductory Exercises into the Renovated Astronomy”) with a view to publication. He continued his study of Brahe’s observational and reduction techniques well into his retirement.

Michael was a generous colleague, always ready to share his insights and experience in scientific and procedural discussions, and a frequent visitor to STScI. He was also a passionate European: a founding member of the European Astronomical Society and a lifetime member as soon as the opportunity arose. He was a low-noise and no-frills person but with a fine sense of humour. He leaves a wife, who also obtained a Ph.D. in astronomy, and a son and daughter.

Dr. Rosa had been a member of the Astronomical Society since 1982 - and thus for 40 years.