Investigating the diurnal and spatial variability of flows in the atmospheric boundary layer: A large eddy simulation study.

摘要

Large-eddy simulation (LES) studies of the atmospheric boundary layer (ABL) have historically modeled the daytime (convective), nighttime (stable) and dawn/dusk windy (neutral) regimes separately under the assumption of a quasi-steady ABL. The real-world ABL however, continuously transitions between the different stability regimes and development of an LES capable of simulating the entire diurnal evolution of the ABL is needed. We have developed an LES tool (The JHU-LES code) with the new-generation Lagrangian dynamic models capable of dynamic adjustment of the subgrid-scale stresses thereby, making it apt for LES over entire diurnal cycles of the ABL. Preliminary LES studies demonstrate that the JHU-LES code reproduces well-known features of the quasi-steady convective and stable boundary layers, such as the well-known spectral scalings for production and inertial subranges. LES of the entire 24-hour diurnal evolution of the atmospheric boundary layer is then performed and compared successfully to field observations (HATS dataset). Important features of the diurnal ABL such as entrainment-based growth of the CBL, development of the stable boundary layer and evolution of the nocturnal low-level jet are well reproduced. The advantages of using a local Obukhov length-scale to normalize the results are highlighted. To investigate the role of surface boundary conditions and geostrophic wind forcing, LES investigations of multi-day evolution of the ABL flow are then performed with several combinations of surface boundary conditions (imposed temperature and heat flux) and geostrophic forcing (constant, time-varying, time and height varying). The variable geostrophic forcing significantly improves the agreement of LES results with surface flux observations but shows poor agreement with daytime surface fluxes and, daytime and nighttime mean profiles. The LES setup using an imposed surface temperature almost always yields better results than cases where the heat flux is imposed. Additional simulations over a variety of heterogeneous surfaces derived from remote-sensed data are performed to assess the role of surface flux boundary condition in flows over heterogeneous terrain. Simulations over surfaces with the same spatial mean but different spatial structure yield different mean temperature and flux profiles with imposed temperature boundary condition. Reynolds decomposition is then applied to identify the relevant terms in the Monin-Obukhov similarity formulation for surface heat flux, leading to a truncated equation with 6 terms mainly comprising spatial covariances of variables such as u*,Ts etc. The results indicate that the impact of surface heterogeneity arises from complex spatial correlations between the imposed properties and the flow variables. We conclude that the full Monin-Obukhov similarity formulation for estimating surface heat flux with an imposed surface temperature should be used as the surface boundary condition for flows over both homogeneous and heterogeneous terrain.