Sensitivity of Deep Convective Updraft Magnitudes to Horizontal Grid Spacing: From Kilometer Scales to LES Scales

Convective cloud updrafts flux mass vertically throughout the atmosphere and have significant impacts on many atmospheric phenomena, including precipitation production, tropospheric and stratospheric composition, and regional and global circulations. As such, it is important to understand the processes that govern convective cloud updrafts and the scales at which they operate. As part of the NASA Investigation of Convective Updrafts (INCUS) mission, over 60 large-domain, high-resolution simulations have been conducted for convective cloud cases using the Regional Atmospheric Modeling System (2-moment RAMS microphysics) and Weather Research and Forecasting model (2-moment Morrison and Thompson microphysics) in support of the development of the INCUS algorithm and scientific approach. The cases span a wide range of convective storm morphologies from isolated deep convective clouds to mesoscale convective systems and tropical cyclones. The simulations utilize three one-way nested domains with horizontal grid spacings of 1.6 km, 400 m, and 100 m, respectively.  

Using the INCUS simulation database, we address the following questions: how much do updraft magnitudes vary as a function of horizontal grid spacing and why? To address these questions, we quantify the differences in vertical velocity between our simulations with 1.6 km, 400 m, and 100 m grid spacing. Initial results show that vertical velocities tend to be stronger below ~ 5 km AGL and weaker above ~5 km AGL, as grid spacing decreases, and that these results are consistent for all three modeling frameworks. We further decompose updrafts into the components of the vertical velocity tendency equation to understand the processes driving vertical velocity differences as a function of grid spacing. With the breadth of the INCUS simulation database, we further quantify how the vertical velocity grid spacing dependency varies as a function of convective system type (e.g. updrafts within scattered convective clouds versus more organized convective systems). This research provides important insights into systematic biases in the representation of deep convective clouds that arise from model horizontal grid spacing and implications for global kilometer-scale modeling.