Kristopher Bedka¹, Jason Brunner², Wayne Feltz², and Richard Dworak²

¹ Science Systems and Applications Inc., Hampton, VA USA
² Cooperative Institute for Meteorological Satellite Studies, Madison, WI USA

An overshooting convective cloud top is defined by the American Meteorological Society as "a domelike protrusion above a cumulonimbus anvil, representing the intrusion of an updraft through its equilibrium level". A single overshooting top (OT) exists for less than 30 mins and has a maximum diameter of ~15 km. Despite their relatively small size and short duration, storms with OTs often produce hazardous weather conditions such as aviation turbulence, frequent lightning, heavy rainfall, large hail, damaging wind, and tornadoes. A five-year OT climatology shows that OTs occur frequently across the continental U.S. and there are clear diurnal differences in OT activity (see Figure 1). Figure 2 illustrates that turbulence is more often observed during aircraft flight near an OT compared to ordinary non-OT cold cloud pixels in the ~11 µm infrared window (IRW) channel. Figure 2 also shows that lightning is more often observed near OTs and that the minimum IRW brightness temperature (BT) within an OT can be used as an indicator of cloud-to-ground lightning activity.

Signatures in multispectral weather satellite imagery indicate the presence of OTs. OTs exhibit a lumpy or "cauliflower" textured appearance in visible channel imagery. The 6 to 7 µm water vapor absorption minus the ~11 µm IRW channel BT difference (WV-IRW BTD) technique for OT detection has been described extensively in the literature (Fritz and Laszlo, 1993; Ackerman, 1996; Schmetz et al, 1997; Setvak et al., 2007; Martin et al., 2008). OTs are also inferred through the presence of a small cluster of very cold brightness temperatures (BTs) in the ~11 µm infrared window (IRW) region. OTs continue to cool at a rate of 7-9 K km^-1 as they ascend into the lower stratosphere (Negri 1982; Adler et al. 1983), making them significantly colder than the surrounding anvil cloud temperature.

Though it is commonly understood that a small cluster of very cold IRW brightness temperatures relates well with the presence of an OT, this characteristic has yet to be exploited in any operational objective OT detection technique. Spatial IRW BT gradients ("IRW-texture" hereafter) can be combined with NWP-based tropopause temperature information and knowledge of the characteristic size of an OT to objectively identify them at their proper scale (Bedka et al. 2009). Such a technique would have some advantages over the WV-IRW BTD in that: 1) it is not explicitly affected by the spatial/vertical distribution of atmospheric water vapor, 2) it does not over-diagnose the size of an individual OT, and 3) it does not use WV BT information which can be affected by variation in the central wavelength and/or spectral coverage of the WV absorption channel.

OTs found in combination with a U or V shaped region of cold infrared window brightness temperatures (BTs) are often indicative of an especially severe thunderstorm. Once OTs have been identified by the IRW-texture technique, the focus can be directed toward the objective detection of the enhanced-V signature. While the enhanced-V is often highly variable in infrared imagery (see Figure 3), one aspect of the enhanced-V remains fairly constant in that the "arms" of the V signature enclose a warm region downwind of the overshooting top to form an "anvil thermal couplet". Brunner et al. (2007) showed that these cold (or enhanced)-U/V producing storms with a minimum IRW BT of = 205 K in the OT region and an anvil thermal couplet of = 7 K produced severe weather for greater than 90% of all events during summers 2003 and 2004. UW-CIMSS and Kristopher Bedka (SSAI/NASA LaRC) have developed a pattern recognition technique with IRW imagery to objectively detect anvil thermal couplets associated with the enhanced-V signature.

These IRW-texture OT and enhanced-V/anvil thermal couplet detection algorithms are currently being developed for future operations with the Geostationary Operational Environmental Satellite Advanced Baseline Imager (GOES-R ABI, Schmit et al. 2005) within the GOES-R Aviation Algorithm Working Group. As GOES-R ABI will offer 2 km spatial resolution in the infrared channels, we can use current satellite instruments to emulate the imagery that will be available in the future with GOES-R ABI. The following discussion provides some examples of algorithm output and validation using MODIS, AVHRR, MSG SEVIRI, CloudSat, and CALIPSO data. For the following examples. MODIS, AVHRR, and MSG SEVIRI imagery have been averaged to the 2 km ABI resolution. The WV-IRW BTD data have been preserved at a 1 km resolution to show the data that is currently available to forecasters.

We must be creative in validating objective OT detection output since a large database of all OT locations throughout the world does not exist,. Figure 4 provides a rare view of an OT producing storm over Mali on 5 February 2008 which was photographed by the International Space Station. MSG SEVIRI imagery shows the characteristic region of cold 10.8 µm IRW BTs and lumpy appearance in 1 km high-resolution visible (HRV) imagery of the OT photographed by the Space Station. The IRW-texture technique identified the OT region perfectly in this case. Another unique way of looking at deep convective storms is through NASA CloudSat and CALIPSO satellite profiles. Figure 5 shows that these satellites passed directly over an OT over the Atlantic Ocean off of the coast of North Carolina. Aqua MODIS IRW and WV BT data and IRW-texture OT detections are co-located with these two satellite profiles to compare IRW-texture and WV-IRW BTD performance. The comparison indicates that the IRW-texture technique again performs well in detecting the ~8 km wide OT. If a 2 K WV-IRW BTD threshold were used here for OT detection, no OT pixels would be detected. If simply a positive BTD were used here, nearly the entire anvil cloud would be detected which would produce a very high false alarm rate. This example, coupled with those shown by Bedka et al. (2009), indicate that the IRW-texture technique offers a more consistent day/night OT detection capability than other existing methods, allowing for unambiguous interpretation and application of product output for aviation and severe weather forecasting. Please see Bedka et al. (2009) or contact kristopher.m.bedka@nasa.gov for a full description of the IRW-texture algorithm and validation in addition to OT relationships with turbulence and cloud-to-ground lightning. In addition, a OT detection climatology over Europe similar to that shown in Figure 1 is also in preparation.

An example of objective enhanced-V/anvil thermal couplet detection is provided in Figure 6. MODIS 1 km IRW imagery from this 7 April 2006 event shows 5 enhanced-V producing storms. OTs and anvil thermal couplets were detected for 4 of the 5 storms. There were no false detections for this case. This detection algorithm was applied to 203 enhanced-V producing storms that occurred across 55 MODIS or AVHRR images. The validation indicates that the probability of enhanced-V detection was 56% and the false alarm rate was 25%. 72% of these 203 storms produced severe weather with +/- 30 mins of the time of the image. 79% of the storms detected by the algorithm were severe and 64% of the undetected storms were severe, indicating that this algorithm is detecting a larger fraction of the severe storms in our database. This validation will be expanded to the 450 enhanced-V storm database described by Brunner et al. (2007). A publication describing this enhanced-V/anvil thermal couplet detection methodology and validation is currently in preparation.

References

Ackerman, S.A., 1996: Global satellite observations of negative brightness temperature differences between 11 and 6.7 µm. J. Atmos. Sci., 53, 2803–2812.

Adler, R. F., M. J. Markus, D. D. Fen, G. Szejwach, and W. E. Shenk, 1983: Thunderstorm top structure observed by aircraft overflights with an infrared radiometer. J. Appl. Meteor. And Climatol., 22, 579-593.

Bedka, K.M., J.C. Brunner, R. Dworak, W.F. Feltz, J.A. Otkin, and T. Greenwald, 2009: Objective Satellite-Based Overshooting Top Detection Using Infrared Window Channel Brightness Temperature Gradients. Accepted for publication in J. Appl. Meteor. and Climatol. (August 2009)

Brunner, J.C., S.A. Ackerman, A.S. Bachmeier, and R.M. Rabin, 2007: A Quantitative Analysis of the Enhanced-V Feature in Relation to Severe Weather. Wea. Forecasting, 22, 853–872.

Fritz, S., and I. Laszlo, 1993: Detection of water vapor in the stratosphere over very high clouds in the tropics. J. Geophys. Res., 98 (D12), 22959-22967.
Martin, D. W., R. A. Kohrs, F. R. Mosher, C. M. Medaglia, and C. Adamo, 2008: Over-ocean validation of the Global Convective Diagnostic. J. Appl. Meteor. Climatol., 47, 525-543.

Negri, A. J., 1982: Cloud-top structure of tornadic storms on 10 April 1979 from rapid scan and stereo satellite observations. Bull. Amer. Meteor. Soc., 63, 1151-1159.

Schmetz, J., S. A. Tjemkes, M. Gube, and L. van de Berg, 1997: Monitoring deep convection and convective overshooting with METEOSAT. Adv. Space. Res., 19, 433-441.

Schmit, T.J., M.M. Gunshor, W.P. Menzel, J.J. Gurka, J. Li, and A.S. Bachmeier, 2005: Introducing the next-generation Advanced Baseline Imager on GOES-R. Bull. Amer. Meteor. Soc., 86, 1079–1096.

Setvak, M., R. M. Rabin, and P. K. Wang, 2007: Contribution of the MODIS instrument to observations of deep convective storms and stratospheric moisture detection in GOES and MSG imagery. Atmos. Res., 83, 505-518.

Figure 1: a) A climatology of overshooting tops detected by the IRW-texture technique using GOES-12 imagery from April-September 2004-2008. b) OT detections separated by the time of occurrence. Warmer colors indicate that OTs were detected more often during the daytime (9 AM - 9 PM Local Time) and cool colors indicate greater OT activity during the night.

Figure 2: (a) The frequency of turbulence inferred through objective EDR turbulence observations from United Airlines aircraft at varying distance from GOES-12 OT and non-overshooting cold pixels (non-OT). This study was performed using data from April-September 2005-2008 over the domain shown in Figure 1. The frequency of severe turbulence is multiplied by 10 so that variability in the curves can be seen using the y-axis scale appropriate for lesser intensity turbulence. (b) (left) The distance between both GOES-12 IRW-texture OT detections and non-overshooting cold pixels to the closest cloud-to-ground lightning strike from May to September 2008. (right) A similar comparison to the left panel, but overshooting and non-overshooting pixels are grouped into IRW BT bins.

Figure 3: An example of 8 cold (or enhanced)-U/V producing storms in MODIS and AVHRR ~11 µm brightness temperature imagery across the continental U.S. The color enhancement and horizontal distance shown in each panel is identical, illustrating the variability of the enhanced-V signature.

Figure 4: (top) A photograph of an overshooting top producing storm over Mali (western Africa) on 5 February 2008. (bottom panels) (left) 3 km MSG SEVIRI 10.8 µm brightness temperature imagery, 1 km SEVIRI high-resolution visible with (right) and without (center) IRW-texture overshooting top detections.

Figure 5: (left) Aqua MODIS 1 km 10.7 µm brightness temperature imagery with IRW-texture OT detections (white dots). (right) IRW-texture OT detections co-located with MODIS brightness temperatures, CloudSat radar reflectivity, CALIPSO cloud top height, and the NASA GEOS-5 model tropopause height analysis.

Figure 6: Aqua MODIS 1 km 10.7 µm brightness temperature imagery of a set of enhanced-V producing storms that occurred on 7 April 2006 at 1845 UTC. Five enhanced-V signatures are outlined with a black dashed line. Overshooting top detections are shown with blue squares and anvil thermal couplet detections are shown with green squares.