Modelling Secondary Circulation in Convective Boundary Layer Using Large Eddy Simulation

Abstract

Mesoscale secondary circulations, which frequently arise over heterogeneous land surfaces, profoundly influence atmospheric structure and characteristics. Its structure is primarily shaped by the combined effects of wind shear and buoyancy. The turbulence generated by the secondary circulation can considerably influence the fluxes, including estimations employed in the Monin–Obukhov similarity theory and measurements in eddy covariance systems. An ever-increasing body of evidence points to secondary circulations as the primary source of the reported underestimation of heat flux (i.e., flux imbalance, FI) by 10% to 30% across various sites. <br /> A series of large eddy simulations (LES) were conducted in this PhD work to investigate the formation of secondary circulations under different conditions and to quantify its impact on flux estimations. These included one-dimensional strip-like soil moisture distribution with ambient wind speeds ranging from 0.5 m/s to 16 m/s in various wind directions (Chapter 3), two-dimensional checkerboard soil moisture distribution with heterogeneous scales varying from 50 m to 2,400 m (Chapter 4). A secondary circulation strength metric is proposed and found to have a positive correlation with the Bowen ratio and heterogeneity parameter, and a negative correlation with wind speeds when the wind direction is perpendicular to the direction of heterogeneity. It was observed that as the strength of the secondary circulation increased, the turbulent heat flux decreased, maintaining the same soil moisture conditions. Two distinct secondary circulation schemes are identified based on the heterogeneity scale: thermally-induced secondary circulations (TMCs) and turbulent organized structures (TOS). <br /> The results of the LES were used to evaluate four selected FI prediction models. These models demonstrated an ability to capture the FI accurately. A novel first-order nonlocal turbulence closure scheme has been proposed (Chapter 5), namely the flux imbalance and K-theory (FLIMK), which employs the FI prediction model to account for the nonlocal processes and the conventional K-theory for the local processes. The FLIMK scheme has been demonstrated to reduce the flux imbalance from 15% to 6% in LES and from 16% to 6.7% in numerical weather prediction (NWP) models.