Internal gravity wave detection in high-resolution model data

Internal gravity waves (GWs) are ubiquitous in the atmosphere, affecting momentum and energy budgets. However, our understanding of GW effects is still incomplete. As they act on various spatial and temporal scales, global observations of GWs face several difficulties and their parametrizations in climate models employ numerous simplifications and are only poorly constrained. Also, GW analyses in high-resolution datasets contain some uncertainty that we aim to quantify and minimize in our research. We study the uncertainty for a Gaussian high-pass filter method applied on a WRF simulation with horizontal resolution of 3 km covering a domain around the Drake Passage and ranging up to the altitude of 80 km. We show that the momentum flux and drag estimates evaluated by the filtering method are sensitive to the value of a cut-off parameter, especially the horizontal drag components. This motivates us to formulate a new, modified filtering method for GW detection that sets an optimal value of the cut-off parameter at each step based on the spectral information – the method uses a wavelength identified in the horizontal spectrum of kinetic energy. Finally, we note that the type of a response function in the high-pass filter definition also impacts the resulting estimates.