Unclouding the Correlations: A Principal Component Analysis of Convective Environments

There are many unknowns surrounding why convective storms occur exactly when and where they do, including environmental impacts on storm morphology and processes. In this study, we first evaluate how the depth and width of precipitating convective storms observed by the TRMM and GPM spaceborne radars are related to large-scale environmental variables obtained from ERA5 reanalysis. Using random forests, we identify eight variables that are useful for distinguishing between global convective storm modes. Furthermore, principal component analysis can collapse most of the information in these variables into just two dimensions. Next, we conduct a lifecycle analysis of convective systems observed and tracked using geostationary satellite observations. We explore the relationships between a system’s initial environment and its longevity, size, and height. We find that the deepest convective systems are observed in environments significantly different from the environments of the widest, most heavily precipitating convective systems, and that more intense storms tend to be associated with initial environments with higher precipitable water. When applied to observations from NASA’s upcoming INvestigation of Convective UpdraftS (INCUS) satellite mission, this method will allow us to quantify relationships between the environment and convective mass flux, thereby addressing one of INCUS’ primary science objectives.