The potential of a ground-based microwave temperature profiler to combine full tropospheric profiling with highresolution profiling of the boundary layer is investigated. For that purpose, statistical retrieval algorithms that incorporate observations from different elevation angles and frequencies are derived from long-term radiosonde data. A simulation study shows the potential to significantly improve the retrieval performance in the lowest kilometer by combining angular information from relatively opaque channels with zenith-only information from more transparent channels. Observations by a state-of-the-art radiometer employed during the International Lindenberg Campaign for Assessment of Humidity and Cloud Profiling Systems and Its Impact on High-Resolution Modeling (LAUNCH) in Lindenberg, Germany, are used for an experimental evaluation with observations from a 99-m mast and radiosondes. The comparison not only reveals the high accuracy achieved by combining angular and spectral observations (overall, less than 1 K below 1.5 km), but also emphasizes the need for a realistic description of radiometer noise within the algorithm. The capability of the profiler to observe the height and strength of low-level temperature inversions is highlighted.