资源环境科技发展态势分析平台

Idealized simulations are used to determine the sensitivity of moist orographic convection to horizontal grid spacing Δ_h. In simulated mechanically (MECH) and thermally (THERM) forced convection over an isolated ridge, Δ_h is varied systematically over both the deep-convection (Δ_h ~ 10–1 km) and turbulence (Δ_h ~ 1 km–100 m) gray zones. To aid physical interpretation, a new parcel-based bulk entrainment/detrainment diagnosis for horizontally heterogeneous flows is developed. Within the deep-convection gray zone, the Δ_h sensitivity is dominated by differences in parameterized versus explicit convection; the former initiates convection too far upstream of the ridge (MECH) and too early in the diurnal heating cycle (THERM). These errors stem in part from a large underprediction of parameterized entrainment and detrainment. Within the turbulence gray zone, sensitivities to Δ_h arise from the representation of both subcloud- and cloud-layer turbulence. As Δ_h is decreased, MECH exhibits stronger cloud-layer entrainment to enhance the convective mass flux M_co, while THERM exhibits stronger detrainment to suppress M_co and delay convection initiation. The latter is reinforced by increased subcloud turbulence at smaller Δ_h, which leads to drying and diffusion of the central updraft responsible for initiating moist convection. Numerical convergence to a robust solution occurs only in THERM, which develops a fully turbulent flow with a resolved inertial subrange (for Δ_h ≤ 250 m). In MECH, by contrast, turbulent transition occurs within the orographic cloud, the details of which depend on both physical location and Δ_h.