An Evaluation of Satellite-Derived Atmospheric Motion Vector (AMV) Characteristics in Tropical Cyclones Using TCI HDSS Dropsondes

Atmospheric motion vectors (AMVs) have been routinely generated and assimilated into numerical weather models for over two decades. The AMVs are assigned a speed, direction, and height based on a series of tracking and filtering algorithms, but the height assignment has been difficult to accurately validate and improve. The problem is exaggerated in the unique and dynamic tropical cyclone outflow environment.

A few previous studies compared AMVs to dropsonde data in tropical cyclone environments, but they did not have the coverage or volume of dropsonde data especially in the upper-levels of the core region. The 2015 ONR Tropical Cyclone Intensity (TCI) field program flew a high-altitude aircraft over the top of TCs with its High-Density Sounding System (HDSS). The higher density of dropsonde data allows stricter quality and matching criteria with the AMVs to be applied, increasing the robustness of the comparisons. And the dropsonde wind profiles collected at very high altitudes over the top of TCs allows for unprecedented comparisons with AMVs in that dynamic region.

Hurricanes Joaquin and Patricia from the 2015 season were thoroughly sampled by the HDSS dropsondes during the TCI field program. Comparisons of AMV-sonde vector difference metrics reveal quality characteristics of the AMVs in the vicinity of TC cores, and are used to determine the discrete level of best fit height assignments for the AMVs. Using the dropsonde wind profiles, we also explore the thickness of the tropospheric layer that the AMVs best represent.