Monitoring Eyewall Convection Using a Bispectral Geostationary Satellite Product

Results from the last decade have suggested a link between the frequent occurrence of convective towers in the eyewall and rapid intensification (e.g., Reasor et al. 2009; Zhang and Chen 2012). These tropopause-penetrating updrafts may trigger subsidence of stratospheric air with extremely high potential temperature and thus rapidly enhance the upper-level warm core. Monitoring such convective towers beneath cold canopies is difficult with currently available remote sensing products. Geostationary satellite instruments are limited to cloud-top observations, but provide excellent spatial and temporal coverage. Microwave imagers on low-earth orbiting platforms provide valuable data from below the obscuring cloud tops, but are irregular and sparse in their coverage. The subject of this study is an easily derived geostationary satellite product, hereafter referred to as the IRWV product, which may offer a way to mitigate the weaknesses of the current remote sensing suite in monitoring convection in TCs, as proposed by Olander and Velden (2009). The IRWV value is the difference of the IR equivalent brightness temperature minus that of the WV channel. It ranges from large positive (~40 K) for clear skies to small positive or zero for high clouds. For updrafts penetrating the tropopause, however, water vapor detrained or pushed into the stratosphere radiates at a warmer temperature than the overshooting cloud top by a few K, causing the IRWV signal to change signs from positive to negative. IRWV imagery offers the advantage of being available in near real-time with regular observations and near hemispheric spatial coverage. A case study of the RI of Hurricane Guillermo (1997) will be discussed to highlight the benefits and pitfalls of the IRWV signal. Strong IRWV < 0 signals are found along sharp north-south gradients of IR brightness temperature that are not associated with overshooting convection. Differential resolution between the IR and WV channels (~4 km and ~8 km, respectively) and possible geo-registration errors also may contribute to false IRWV < 0 signals. However, a convective burst is well identified and tracked using IRWV in Guillermo. Cooling tops as seen in IR suggest the burst is intensifying as it approaches the upshear quadrant of the storm, where it should be beginning to decay. On the other hand, the negative IRWV signals are strongest around the origin of the burst in the downshear-left sector and lose intensity as the burst moves upshear. The locations of the IRWV < 0 signals match up well qualitatively with the coldest IR tops and “bubbling” areas seen in visible imagery. Super Rapid Scan Operations (SRSO) imagery with periods of ~1 minute resolution during Guillermo's RI reveal much greater temporal variability in the IRWV signal than in the smoothly changing IR. It remains to be seen whether this variability truly reflects changes in the supporting updrafts or is an artifact of the IRWV product. Newer iterations of the GOES imager now have identical resolution for the IR and WV channels, which may reduce some of the blatantly false IRWV < 0 signals evident in the imagery presented here. Ongoing work is examining whether this improvement leads to cleaner and more useful IRWV imagery.