Automated OT detection methods have been developed in recent years at NASA Langley Research Center using support primarily from the NOAA GOES-R program. OTs appear anomalously cold in satellite infrared (IR) imagery and highly textured in visible imagery, signatures that can be reliably detected using pattern recognition methods. Imagery-based pattern recognition is combined with 1) NWP tropopause and equilibrium level analyses and 2) a large sample of both OT and non-OT anvil regions identified within 0.25 km MODIS imagery to develop a model to statistically differentiate between the two populations and arrive at an OT Probability product. Co-located high OT Probability values and visible texture detection are quite reliable indicators that an OT is truly present.
The OT detection products have been designed to process imagery from almost any high spatial resolution (~4 km IR and ~1 km visible at satellite nadir) geostationary or polar-orbiting imager in a very efficient manner, enabling one to acquire and process almost the entirety of a GOES Northern Hemisphere scan (~8000x18000 visible pixels) in ~3 minutes. The algorithm's flexibility and speed has enabled product usage in a wide variety of weather and climate applications. These include 1) generation of a 20-year GOES-East and -West OT database over much of the Western Hemisphere, 2) analysis of hail storm risk over Europe, 3) analysis of climatological OT distributions over the Lake Victoria region, 4) identification of the source regions of enhanced stratospheric water vapor observed by instruments aboard the NASA ER-2 research aircraft, 5) development of a satellite-based aircraft engine icing (i.e. "high ice water content") probability product, 6) analysis of severe storms using 1-minute resolution "super rapid scan" satellite imagery, and 7) near real-time OT product generation over the U.S. This presentation will highlight many of these recent weather and climate applications of NASA OT detection products.