Spatial and temporal analysis of high-frequency internal gravity wave signatures in brightness temperature satellite images

Highly idealized model studies suggest that convectively generated internal gravity waves in the troposphere with horizontal wavelengths on the order of a few kilometers may affect the lifetime, spacing, and depth of clouds and convection. To answer whether such a convection-wave coupling occurs in the real atmosphere, one needs to find corresponding events in observations. In general, the study of high-frequency internal gravity wave-related phenomena in the troposphere is a challenging task because they are usually small-scale and intermittent. To overcome case-by-case studies, it is desirable to have an automatic method to analyze as much data as possible and provide enough independent and diverse evidence.
Here, we focus on brightness temperature satellite images, in particular so-called satellite water vapor channels. These channels measure the radiation at wavelengths corresponding to the energy emitted by water vapor and provide cloud-independent observations of internal gravity waves, in contrast to visible and other infrared satellite channels where one relies on the wave impacts on clouds. In addition, since these water vapor channels are sensitive to certain vertical layers in the troposphere, combining the images also reveals some vertical structure of the observed waves.
We propose an algorithm based on local Fourier analyses to extract information about high-frequency wave patterns in given brightness temperature images. This method allows automatic detection and analysis of many wave patterns in a given domain at once, resulting in a climatology that provides an initial observational basis for further research. Using data from the instrument ABI on board the satellite GOES-16 during the field campaign EUREC⁴A, we demonstrate the capabilities and limitations of the method. Furthermore, we present the respective climatology of the detected waves and discuss approaches based on this to address the initial question.