The effect of unsteady and baroclinic forcing on predicted wind profiles in Large Eddy Simulations: Two case studies of the daytime atmospheric boundary layer
Pedersen, Jesper Grønnegaard; Kelly, Mark; Gryning, Sven-Erik; Brümmer, Burghard
Due to its fine-resolution requirement and subsequent computational demand, Large Eddy Simulation of the atmospheric boundary layer is limited in most cases to computational domains extending only a few kilometers in both the vertical and horizontal directions. Variations in the flow and in relevant atmospheric fields (e.g. temperature) that occur at larger scales must be imposed through boundary conditions or as external forcing. In this work we study the influence of such variations on the wind profile in Large Eddy Simulations of daytime atmospheric boundary layers, by comparing observations with simulations that use progressively more realistic forcing relative to observed large-scale pressure gradients.Two case studies are presented. One is based on measurements from the rural site of Høvsøre in Denmark, and the other on measurements from a suburban site in Hamburg, Germany. The applied domain-scale pressure gradient and its height- and time-dependence are estimated from LIDAR measurements of the wind speed above the atmospheric boundary layer in the Høvsøre case, and from radio soundings and a network of ground-based pressure sensors in the Hamburg case.In the two case studies, LIDAR measurements of the wind speed up to heights between 900 and 1600 m and tower-based measurements up to 100 and 250 m are used to evaluate the performance of the variably-driven Large Eddy Simulations. We find in both case studies that including height- and time-variations in the applied pressure gradient has a significant influence on simulated wind speeds, and improves agreement with measured wind speeds, especially in the Høvsøre case. In the Hamburg case, an overly simplified specification of the height dependence of the forcing, as well as the influence of phenomena such as large-scale subsidence and advection, tend to reduce agreement with measurements, relative to the Høvsøre case. The Hamburg case illustrates that measurements of the surface pressure gradient and relatively infrequent radio soundings alone are not sufficient for accurate estimation of a height- and time-dependent pressure gradient.