Different electromagnetic reflection methods can operate from satellites, airplanes or ground vehicles to illuminate the surface and the inside of ice sheets across varying spatial scales. The backscattered signal is formed by micro-physical ice properties, many of which in turn are influenced by mechanisms operating on a macro-scale: the alignment of crystal orientation fabric (COF) depends on the specific strain regime and the initial impurity loading; the pattern in internal layering is imprinted by accumulation and the surrounding flow regime; the brightness of the bottom reflection over ice-shelves depends on the melting or refreezing of platelet ice which is susceptible to changing ocean currents and grounding line positions. The study is subdivided in five chapters which have been or will be published separately. All studies focus on the link between these small-scale features and their large-scale expressions. The synthesis of the different methods improves the capability of remote sensing to deliver variables from which the current state of the cryosphere can be determined. The study contributes to assimilating geophysical data into the coming generation of ice-dynamic models which improve the understanding of ice-sheet histories and prognose their future behaviour. The starting point is the appearance of the radio-echo free zone (EFZ), which is a feature-less band observed above the ice–bed interface in many radargrams across Greenland and Antarctica. The comparison of the EFZ onset with optical ice-core images yields a connection between the mm-cm scale disturbances in the core’s stratigraphy and the disappearance of radar reflection horizons. This is evidence that ice flow can disturb internal ice layering, which hampers the derivation of a coherent age–depth scale and indicates a changing flow behaviour with depth. A polarimetric radar survey in the same study area shows that backscattered power varies with the horizontal orientation of the antennas (i.e. with the polarization plane). Extrema in backscatter for a constant antenna angle change in direction at 900 m depth. By using different scattering models I differentiate between competing mechanisms for the observed anisotropy, namely a vertically varying COF or ellipsoidal air bubbles. The analysis suggests that the effect from a varying COF is superior. Radar polarimetry thus is capable to infer the principal components of COF, and with it the principal components of the corresponding stress–strain regime. Potentially, the different modes in COF variations are related to climate signals. The detection of anisotropic COF is an important factor in terms of anisotropic ice flow, which in turn influences the shape in internal layering especially near ice divides. The third study combines satellite data and GPS measurements with airborne and ground-based radar surveys to characterize a potential drill site with respect to internal layering, accumulation, and the topography of surface and bedrock. Particular attention is given to the upwarping of internal layering beneath the divides, which is a consequence of a nonlinear and anisotropic rheology. All datasets are used as input for a two-dimensional, anisotropic flow model which estimates an age–depth distribution. Likely, ice from the last glacial is present at larger depths. Ice at intermediate depths seems suitable for studies targeting the last few thousand years. Based on the structure of internal layering, the ice divides and the surrounding flow regime were stable over the last 10 000 years. The investigated ice ridge is surrounded by ice shelves. Surface velocities as well as the transition of grounded and floating ice are mapped from satellite images. This is used in the last two studies. In a collaborative project, the most landward freely floating line of ice shelves, and the most seaward line where ice flow is still influenced by the bedrock are mapped from optical satellite imagery around the Antarctic perimeter. The simultaneous derivation of coastal elevation as a basis for the derivation of flow velocities is a starting point for mass balance estimates, in which ice flow, but also sub-ice shelf melting are important parameters. The latter is treated in a case study. Using the continuity equation in a steady-state approximation, it is possible to spatially map the sub-ice shelf melt rates. A new approach is presented which circumnavigates the interpolation of ice-thickness data by applying the continuity equation only along profile lines.