Convectively generated gravity waves simulated by NAM‐SCA

Convectively generated gravity waves are an unresolved process in global models. However, parameterization of their effects is difficult because the parameterized wave spectrum must be linked to the convective source, which is itself parameterized. Here we use the NAM‐SCA model (nonhydrostatic anelastic model with segmentally constant approximation), which is midway between a cloud‐resolving model and convection parameterization. Unlike conventional convection parameterizations, NAM‐SCA is built directly upon a nonhydrostatic anelastic model, and thus it describes dynamical processes consistently but in a “compressed” manner (i.e., with truncation via a spatiotemporal adaptive mesh) that facilitates its potential use as a parameterization. As a result, under the dynamical framework of NAM‐SCA, convection generates a spectrum of gravity waves in a physically consistent manner without any additional assumptions or simplifications. The present study examines the capacity of NAM‐SCA for generating gravity waves from convection under varying degree of “compression” and assesses its viability as a physically consistent convective gravity wave source parameterization. By taking the TWP‐ICE (Tropical Warm Pool‐International Cloud Experiment) period for the case study, NAM‐SCA successfully represents the convectively generated gravity waves even down to the compression rate of 0.1. Analysis in the wave number frequency space shows that the compression behaves as a weak low‐pass filter in wave number, but higher‐frequency components generated by dynamic adaptation partially compensate the filtering effect making the net Reynolds stress relatively insensitive to compression.