Comparison of temperature and wind profiles between ground-based remote sensing observations and numerical weather prediction model in complex Alpine topography: the Meiringen campaign

Abstract

Thermally driven valley winds and near-surface air temperature inversions are common in complex topography and have a significant impact on the local and mesoscale weather situation. They affect both the dynamics of air masses and the concentration of pollutants. Valley winds affect them by favoring horizontal transport and exchange between the boundary layer and the free troposphere, whereas temperature inversion concentrates pollutants in cold stable surface layers. The complex interactions that lead to the observed weather patterns are challenging for numerical weather prediction (NWP) models. To study the performance of the COSMO-1E (Consortium for Small-scale Modeling) analysis, which is called KENDA-1 (Km-Scale Ensemble-Based Data Assimilation), a measurement campaign took place from October 2021 to August 2022 in the 1.5 km wide Swiss Alpine valley of the Haslital. A microwave radiometer and a Doppler wind lidar were installed at Meiringen, in addition to numerous automatic ground measurement stations recording meteorological surface variables. Near the measurement site, the low-altitude Brünig Pass influences the wind dynamics similarly to a tributary. The data collected show frequent nighttime temperature inversions for all the months under study, which persist during the day in the colder months. An extended thermal wind system was also observed during the campaign, except in December and January, allowing for an extended analysis of the winds along and across the valley. The comparison between the observations and the KENDA-1 data provides good model performance for monthly temperature and wind medians but frequent and important differences for single profiles, especially in the case of particular events such as foehn events. Modeled nighttime ground temperature overestimation is common due to missed temperature inversions, resulting in a bias of up to 8 °C. Concerning the valley wind system, modeled flows are similar to the observations in their extent and strength but suffer from too early a morning transition time towards up-valley winds. The findings of the present study mostly based on monthly averages allow for a better understanding of the temperature distributions, the thermally driven wind system in a medium-sized valley, the interactions with tributary valley flows, and the performance and limitations of KENDA-1 in such a complex topography.