Profiling the Arctic Stable Boundary Layer in Advent Valley, Svalbard: Measurements and Simulations

The unmanned aerial system SUMO (Small Unmanned Meteorological Observer) has been used for the observation of the structure and behaviour of the atmospheric boundary layer above the Advent Valley, Svalbard during a two-week period in early spring 2009. Temperature, humidity and wind profiles measured by the SUMO system have been compared with measurements of a small tethered balloon system that was operated simultaneously. It is shown that both systems complement each other. Above 200 m, the SUMO system outperforms the tethered balloon in terms of flexibility and the ability to penetrate strong inversion layers of the Arctic boundary layer. Below that level, the tethered balloon system provides atmospheric profiles with higher accuracy, mainly due to its ability to operate at very low vertical velocities. For the observational period, a numerical mesoscale model has been run at high resolution and evaluated with SUMO profiles reaching up to a height of 1500 m above the ground. The sensitivity to the choice of atmospheric boundary-layer schemes and horizontal resolution has been investigated. A new scheme especially suited for stable conditions slightly improves the temperature forecast in stable conditions, although all schemes show a warm bias close to the surface and a cold bias above the atmospheric boundary layer. During one cold and cloudless night, the SUMO system could be operated nearly continuously (every 30-45 minutes). This allowed for a detailed case study addressing the structure and behaviour of the air column within and above Advent Valley and its interaction with the local topography. The SUMO measurements in conjunction with a 10-m meteorological mast enabled the identification of a very stable nocturnal surface layer adjacent to the valley bottom, a stable air column in the valley and a strong inversion layer above the summit height. The results indicate the presence of inertial-gravity waves during the night, a feature not captured by the model.