<p>Microwave radiometers can provide continuous temperature and humidity profiles in the troposphere. This contribution presents data from the long-term deployment (2012–2025) of a passive microwave radiometer in the Inn Valley, Austria. In a first step, an automated quality control procedure was developed and applied to the dataset, resulting in over 42,000&#xa0;h of data for analysis. Comparison against radiosoundings shows that the microwave radiometer captures the boundary layer temperature structure reliably up to about 1500&#xa0;m above ground level, with root-mean-square errors (RMSE) of 0.5–1.0&#xa0;K. At higher altitudes errors increase and a positive temperature bias of up to 1.4&#xa0;K emerges. Absolute humidity retrievals perform less well overall with RMSE values around 0.6–0.9&#xa0;g m^− 3 in the lowest 1500&#xa0;m. The microwave radiometer is unable to capture sharp gradients in temperature (or humidity) and thus underestimates positive lapse rates and overestimates negative lapse rates. Next, this long-term dataset is used to investigate the seasonal and diurnal variability of the mountain boundary layer. Analysis of normalised virtual potential temperature profiles highlights the complexity of the valley atmosphere, which frequently deviates from the textbook structure occurring over flat and homogeneous terrain. Above the surface layer, in the warm season a mixed layer develops with a stably stratified layer above, sometimes accompanied by an elevated neutral layer below crest height. The first mixed layer is shallower than expected from studies in flat terrain under comparable conditions. At night, the residual layer, which is typically present over flat terrain, is rarely seen in this dataset. Considerable differences in boundary layer structure during valley-wind days and foehn events are highlighted. Overall, this study demonstrates both the potential and limitations of passive microwave radiometry measurements for boundary layer research in complex terrain.&#xa0;</p>