A technique to identify vortex air using carbon monoxide observations

This study uses observations of carbon monoxide (CO) from the Microwave Limb Sounder instrument to identify vortex air in the winter polar regions. In particular, we use the probability distribution function (PDF) of the CO data to delineate CO concentrations characteristic of the interior of the vortex core as a function of space and time. This is achieved by fitting two Gaussian distributions to the PDF for a specific period and vertical level. These Gaussian fits are then examined to determine whether two chemically distinct regions exist by inspecting the intersection area between the Gaussians relative to their individual areas. When chemically distinct regions exist, the values of the fitted mean CO concentrations are representative of the interior and exterior of the polar vortex. To prove this point, a domain-filling analysis is performed to produce high-resolution maps. Comparison of these maps with the vortex edge derived using the equivalent latitude technique, and a statistical analysis over a range of years and isentropic levels, shows that the fitting scheme can be used to characterize measurements made inside the vortex during periods when chemically distinct populations exist. The statistical analysis also suggests that a threshold CO concentration can be derived based on the Gaussian fits which defines a good mixture between reducing the quantity of CO observations incorrectly identified as vortex interior air, whilst also minimizing the number of observations misclassified as vortex exterior air. This technique therefore provides an alternative to the dynamically derived calculation of values characteristic of the vortex interior.