Statistical relationships between satellite-derived mesoscale atmospheric motion vectors, rawinsondes, and NOAA wind profiler network observations

High-density, multi-spectral, satellite-derived wind vectors are being produced in real-time using the UW-CIMSS algorithm. These so-called "mesoscale" atmospheric motion vectors (AMVs) can be used to track cloud and water vapor features across sequences of 15-minute GOES imagery and monitor their time evolution. When this process is applied to convective clouds, one can identify cumulus that are rapidly evolving into precipitating cumulonimbus. Developing convective storms are a known aviation hazard due to their strong vertical motion fields and their propensity to produce vertically and horizontally propagating gravity waves. This convective cloud growth rate product is shown to be quite useful within a convective storm initiation (CI) nowcasting algorithm developed through the NASA-sponsored Advanced Satellite Aviation-weather Products (ASAP) project.

The relative accuracy of these vectors (in relation to a robust operational wind observing system) had not been explored in the development of this CI nowcasting product. Understanding how these AMVs relate to actual wind measurements was somewhat irrelevant, as one could properly track a cumuliform cloud feature across a three image sequence without knowing the true accuracy of these AMVs. We must now begin to understand the relationships between satellite AMVs and an accurate high temporal resolution observing system, as AMVs show great promise for use in applications that demand this information, such as horizontal and vertical wind shear zone identification or data assimilation activities. Wind observations from the NOAA profiler network are directly compared to satellite AMVs in this presentation and the statistical relationships between the two datasets will be shown for an eight-month period, where over 10,000 co-located AMVs and wind profiles (in the horizontal, vertical, and time) were collected.