Magnetic fields of chemically peculiar (CP) stars of the upper main sequence are characterised by a geometry organised at a large scale, permeating the entire photosphere, and with a typical strength of the order of 0.1–30 kG. Here I review the results obtained from statistical studies of the structures of the magnetic fields of CP stars, which are aimed at finding how magnetic strength and morphology are correlated with other stellar characteristics and with the star's evolutionary state.

The dynamical behavior of upper cool-star photospheres constitutes the next step in realistic atmosphere modeling after the successful numerical implementation of low-photosphere granulation. The solar example shows that this behavior is complex even when magnetism is ignored. Acoustic waves are a principal ingredient, but so are atmospheric gravity waves. Acoustic events provide less pistoning than expected. Correlation and Fourier analyses of image sequences from TRACE demonstrate the ubiquity of gravity waves.

Over the past few years there has been growing interest in helical magnetic field structures seen at the solar surface, in coronal mass ejections, as well as in the solar wind. Although there is a great deal of randomness in the data, on average the extended structures are mostly left-handed on the northern hemisphere and right-handed on the southern. Surface field structures are also classified as dextral (= right bearing) and sinistral (= left bearing) occurring preferentially in the northern and southern hemispheres respectively. Of particular interest here is a quantitative measurement of the associated emergence rates of helical structures, which translate to magnetic helicity fluxes. In this review, we give a brief survey of what has been found so far and what is expected based on models. Particular emphasis is put on the scale dependence of the associated fields and an attempt is made to estimate the helicity flux of the mean field vs. fluctuating field.

This paper highlights the increasing interest of astrophysical spectropolarimetry by showing how remote sensing techniques based on the Hanle and Zeeman effects are allowing us to investigate the magnetism of the extended solar atmosphere. Particular emphasis is given to the development of new diagnostic windows on the weakest magnetic fields of the photosphere, chromosphere and corona. Spectropolarimetry in the He i 10830 Å multiplet is allowing us to infer the three-dimensional geometry of the magnetic fields that confine the plasma of solar prominences. Multilevel modelling of the Hanle and Zeeman effects in the Ca II IR triplet and other chromospheric lines is helping us to decipher the strength and topology of chromospheric magnetic fields in regions where photospheric magnetograms show the well-known ‘salt and pepper’ patterns of mixed polarities. The Hanle effect in molecular lines, as well as in atomic lines of rare-earth elements like Ce II, offers a novel diagnostic tool for empirical investigations of ‘turbulent’ magnetic fields in relatively deep regions of the ‘quiet’ solar photosphere.

The review will cover the following topics: (1) Ionization energies; (2) Partition functions; (3) Sources of data for atomic and ionic wavelengths, transition probabilities, and broadening parameters, including nuclear effects (hfs and isotope shifts); (4) Opacities from photoionization of abundant elements (atoms and atomic ions) with emphasis on integration of TOPBASE material; and (5) Data bases for diatomic molecules. We emphasize topics of direct relevance to the synthesis of stellar spectra, primarily within the domain where LTE is useful. Additional parameters, such as line-broadening parameters, or excitation cross sections are not reviewed.

I describe recent progress in terms of 3D hydrodynamical model atmospheres and 3D line formation and their applications to stellar abundance analyses of late-type stars. Such 3D studies remove the free parameters inherent in classical 1D investigations (mixing length parameters, macro- and microturbulence) yet are highly successful in reproducing a large arsenal of observational constraints such as detailed line shapes and asymmetries. Their potential for abundance analyses is illustrated by discussing the derived oxygen abundances in the Sun and in metal-poor stars, where they seem to resolve long-standing problems as well as significantly alter the inferred conclusions.

Models are compared with observations in several examples, including spectroscopic temperature measurement, irradiance variations, stellar rotation from line profiles, and velocity fields in stellar photospheres. A few key issues for investigations in the near future are suggested.

We present spectroscopic and photometric observational evidence for abundance stratification in stellar atmospheres. Attention is given to chemically peculiar (Ap) stars in which magnetic fields stabilize the atmosphere, allowing diffusion processes to establish abundance stratification during the early stages of star's life. The results of recent empirical modelling of chemically stratified atmospheres are briefly discussed, and a comparison is given with the predictions of self-consistent atmospheric models which include radiative diffusion.

The importance of abundance stratification analysis is demonstrated for rapidly oscillating (roAp) stars in which the amplitudes of the radial velocity pulsations are different for the lines of different elements in different ionization stages. We also demonstrate that chemical stratification has important effects on the Zeeman Stokes IQUV profiles of Ap stars, indicating that stratification must be taken into account in detailed modelling of their magnetic fields.

Long-baseline interferometry at optical and near-infrared wavelengths is an emerging technology which is quickly becoming a useful tool to investigate stellar atmospheres and to compare observations with models. Stellar atmosphere models have so far mainly been constrained by comparisons with stellar spectra which are integrated over the stellar disks. Interferometric observations provide spatially and spectrally resolved information and can thus provide important complementary observational information which can be compared to model predictions. Here, I summarize the different aspects on this topic which were discussed at a round table on Thursday, June 20, 2002, during IAU Symposium 210. This summary gives an overview on discussed interferometric facilities and techniques, concepts to study atmospheres by optical interferometry, and particular classes of objects. We conclude that more frequent interactions between the efforts of atmosphere modelling and interferometric observations promise to lead to increased confidence in stellar model atmospheres and to further advancement of the field in the next years.

We started a comparison between different NLTE codes that calculate the statistical equilibrium in solar-type and late type stars (G - A type stars). Discrepancies between different authors analyzing the statistical equilibrium of the same element in the same atmosphere are often surprisingly large. Hence, this study was meant to nail down the origin of these differences. The preliminary results indicate that even if the atmosphere and the atomic input data is fixed, discrepancies of up to 40% in the outer atmospheric layers are still present; the main reason is the different treatment of the background opacity. Following up, we shortly discuss the completeness and accuracy of atomic data used in analyses of the kinetic equilibrium of atoms in the atmospheres of middle and late type stars.

During this round table Bikmaev described the systematic differences among equivalent width measurements and Gulliver discussed the reduction of high dispersion, high signal-to-noise CCD spectra. Smalley considered atmospheric extinction as it affects astronomical measurements while Adelman presented an overview of the ASTRA spectrophotometer.

We summarize the current status of our latest generation of model atmospheres for pulsating asymptotic giant branch stars, discussing effects of non-grey radiative transfer, dust grain properties and drift between gas and dust on the atmospheric structures and wind characteristics. In addition, we give an overview of the resulting synthetic spectra and how they compare with observations.

Milliarcsecond resolution observations of cool stars are becoming increasingly common and sophisticated as recent advances in telescope technology mature. To varying degrees, these observations rely up on stellar models for interpretation of their data, while at the same time present particular challenges to those models. Indications of departures from spherical symmetry are beginning to be observed as increasingly rich image information is obtained by a new generation of interferometers. Examination the subtle variations of wavelength-specific sizes exhibits rich structure, connected to the atmospheric chemistry. For the pulsating stars, such as Mira variables, that structure varies with time, with the phase lags between the various sizes being connected to the atmospheric dynamics. Complex morphologies associated with atmospheric winds have been revealed with these high resolution experiments. A review of these recent results will be presented, concentrated on their implications upon stellar modelling, and the prospects for future observational data.

SiO, H2O and OH masers are ubiquitous in the circumstellar envelopes (CSEs) of oxygen-rich, red giant stars. Radio interferometry allows the imaging of stellar maser emission down to sub-milliarcsecond scales. Such observations reveal asymmetry, inhomogeneity and apparent clumpiness in the extended atmosphere and surrounding envelope of the star. The studies place constraints on processes which are seldom included in models. Here, I review the observational data on stellar masers and discuss their implications for our understanding of the mass-loss and evolution of red giant stars.

The current status of modelling the wind driving mechanism of AGB stars is reviewed. Our present understanding of these winds is substantially guided by spherically symmetric, time-dependent models which reveal a complex interplay between stellar pulsation, hydrodynamics, small-scale waves, chemistry, dust formation and radiative transfer. The level of approximation used for the physical description of the various processes, and the degree of consistence in the models are crucial and different approaches lead to rather different results. Beside reviewing the frontiers of the present theoretical research on AGB star winds, this paper presents some new results concerning the cooling of the gas behind propagating shock waves in the outer atmosphere and the self-organisation of dust-forming gases (cloud formation) caused by a radiative/thermal instability of dust forming gases.

Speakers for this session were selected according to our perceptions about the frontier areas in dynamical atmospheres, the atmospheres of very cool stars, and stellar winds. This summary of posters and topics raised in the round table discussion opens with a brief summary of current issues in modeling pulsating stars and winds. Observational constraints are emphasized in Section 3. on “peeling the onion” of the atmosphere/wind structure. Section 4. covers a variety of methods for detecting and measuring stellar winds. The concluding remarks include some questions facing theorists and observers working in this field.

I offer some comments on the importance of our field (rather modestly, we could say that we are central to almost the entire enterprise of astrophysics) and on some of recent developments in our field (an impressive record of difficult problems attacked and – often – solved). Finally I consider some of the work still left to be done, which is assigned to the participants, their colleagues, friends, and students as homework to be done, as far as possible, for the next major stellar atmosphere meeting.

Electronic versions of the poster contributions are contained on the companion CD-ROM.