12.4 USING C-BAND DUAL-POLARIZATION RADAR SIGNATURES TO IMPROVE CONVECTIVE WIND FORECASTING AT CAPE CANAVERAL AIR FORCE STATION AND NASA KENNEDY SPACE CENTER Corey G. Amiot* and Lawrence D. Carey The University of Alabama in Huntsville, Huntsville, Alabama William P. Roeder and Todd M. McNamara 45th Weather Squadron, Patrick Air Force Base, Florida Richard J. Blakeslee NASA Marshall Space Flight Center, Huntsville, Alabama 1. INTRODUCTION AND BACKGROUND of positive differential reflectivity (Zdr) values above the environmental 0 °C level (Tuttle et al. 1989), and a The United States Air Force’s 45th Weather region of near-0 dB Zdr values extending below the 0 °C Squadron (45WS) is the organization responsible for level within the descending precipitation core of the monitoring atmospheric conditions at Cape Canaveral storm (Wakimoto and Bringi 1988, Scharfenberg 2003). Air Force Station and NASA Kennedy Space Center Additional studies have examined the effects that (CCAFS/KSC) and issuing warnings for hazardous environmental conditions have on precipitation weather conditions when the need arises. One such processes within downburst-producing thunderstorms. warning is issued for convective wind events, for which Srivastava (1987) showed that precipitation ice can be lead times of 30 and 60 minutes are desired for events very important to the formation and intensification of wet with peak wind gusts of 35 knots or greater (i.e., downbursts, especially in warm humid environments. Threshold-1) and 50 knots or greater (i.e., Threshold-2), Meischner et al. (1991) showed that the melting of respectively (Roeder et al. 2014). falling precipitation ice over a shallow layer beneath the 45WS forecasters use a variety of instrumentation 0 °C level can significantly enhance downburst intensity, to increase protection for personnel, facilities, space and White (2015) indicated how Zdr can be used to launches, and space mission payloads at CCAFS/KSC, examine the melting of precipitation ice. including a C-band dual-polarization radar (45WS-WSR) These and additional radar signatures were (Roeder et al. 2009) and the Cape Weather Information explored in this study. Given that the typical lifecycle of Network Display System (Cape WINDS) – a network of ordinary convective cells in the southeastern United 29 weather observation towers strategically placed States is around 20 – 40 minutes (Smith et al. 2004), throughout and around the CCAFS/KSC complex (Fig. attention was given to longer-lived thunderstorm 1). Even with this wealth of technology available, systems (e.g., multicellular convection) in an effort to forecasting convective wind events (e.g., downbursts) is increase lead times for 45WS convective wind warnings. a difficult process. It can be challenging to identify which It was hypothesized that the aforementioned radar storms will produce threshold-level downbursts (i.e., signatures can be observed within the multiple updraft- peak wind gust of 35 knots or greater) from those that downdraft cycles that occur within multicellular will not, and the lead times offered for threshold-level (multicell) convection, which can be used by 45WS events are often shorter than desired by the 45WS. forecasters to identify storm systems that may The purpose of this study is to identify C-band dual- eventually produce a threshold-level downburst at the polarization radar signatures that can be used by 45WS CCAFS/KSC complex. forecasters in real-time to: Preliminary results of this study indicate that certain 1) Provide increased lead times for downburst dual-polarization radar signatures may offer increased events at CCAFS/KSC; lead times for threshold-level downbursts. Specifically, it 2) Decrease false alarm ratios (FARs) for 45WS was found that the peak height of the 1 dB contour convective wind warnings; within a Zdr column, the peak height of the 30 dBZ Zh 3) Differentiate between storm cells that will contour co-located with Zdr values around 0 dB, the produce below-threshold wind gusts, peak Zh value, and the gradient in Zdr below the 0 °C Threshold-1 gusts, and Threshold-2 gusts. level within a descending reflectivity core (DRC) may Past studies have noted various radar signatures that, offer average lead times of 40 – 46 minutes, especially when placed in an environmental context, can provide in multicellular systems. insight into the physical processes occurring within downburst-producing thunderstorms. Some examples 2. DATA AND METHODOLOGY include the peak value of radar reflectivity factor (Zh) within a storm cell (Loconto 2006), a column-like region 2.1 Data ____________________________________________ *Corresponding author address: (insert information All radar data used in this study were from the C- here) band 45WS-WSR, and were provided by the 45WS. (and here) Environmental conditions were analyzed using (and here) atmospheric soundings from the KXMR site at the KSC skid strip. The peak wind gusts from the selected Raw Zh and Zdr values were converted from logarithmic downburst events were observed on the Cape WINDS units to linear units before linear data interpolation was (Fig. 1). The KXMR and Cape WINDS data were applied during the gridding process, before being provided by the 45WS and the United States Air Force’s converted back into logarithmic units for visualization. 14th Weather Squadron. For this study, 10 days with Py-ART was then used to export each gridded threshold-level downbursts were analyzed, which 45WS-WSR volume scan as a Net-CDF file. A separate included 14 threshold-level downbursts and 4 null IDL code was written to read-in and visualize the events (i.e., downbursts with a peak wind gust less than gridded radar data. For each recorded threshold-level 35 knots). All threshold-level and null events occurred wind gust, the storm cell that produced the gust was from May – September 2015, which was the 2015 warm identified using a combination of composite reflectivity, season at CCAFS/KSC. horizontal cross sections, and the 5-minute mean wind direction at the time of the recorded gust. The downburst-producing cell was then manually tracked back in time using composite reflectivity and horizontal cross sections. At several times in the cell’s lifecycle, north-south (N-S) and east-west (E-W) vertical cross sections were taken through multiple locations within the cell to help understand the physical processes occurring within the storm. In cases where storm mergers were observed, all cells within the merger were tracked and analyzed. In multicellular events, all cells earlier in the multicell system that were responsible for forming the downburst-producing cell (e.g., through convection initiated along gust fronts from collapsing cells) were tracked and analyzed. Storms that produced a peak wind gust of 35 knots or greater (i.e., met Threshold-1) were tracked until approximately 50 minutes before the time of the recorded peak gust (if possible), while storms that produced a peak wind gust of 50 knots or greater (i.e., met Threshold-2) were tracked until approximately 80 minutes before the time of the recorded peak gust (if possible). Since the recorded peak wind gust was from a 5-minute observation period in each case, a median time of 2.5 minutes before the Figure 1. Current set up for the Cape Weather time of the reported peak gust was assumed to be the Information Network Display System (Cape WINDS). actual time of the downburst (e.g., a peak gust recorded at 1850 UTC occurred between 1845 – 1850 UTC, and 2.2 Methodology was assumed to have occurred at 1847:30 UTC). The peak wind gust recorded on the Cape WINDS was The first step in this study was to identify all assumed to be the true peak wind gust in each case. threshold-level wind gusts from the Cape WINDS Additional codes written in IDL were used to network during the May – September 2015 time period. examine environmental data, which included calculating A code was written in the Interactive Data Language the 0 °C height, the height of minimum equivalent (IDL) to parse a text file containing all Cape WINDS data and print the date, time, tower ID number, sensor potential temperature (θe), temperature lapse rates, and relative humidity values from the KXMR soundings. For height, sensor direction, 5-minute peak wind gust, and this study, the KXMR sounding nearest the downburst- 5-minute mean wind direction for all Cape WINDS producing system’s lifetime was used in each case. sensors that recorded a wind gust of 35 knots or greater The radar data were analyzed in combination with during this time period. Raw 45WS-WSR data were then the environmental data to better understand physical examined around the time of the recorded threshold- processes occurring within the downburst-producing level gusts using the GR2Analyst software (Gibson 2005) to ensure the presence of a thunderstorm at the storm systems. Common trends observed in Zh, Zdr, and time and location of the recorded wind gust. correlation coefficient (ρhv) were identified using vertical Following that, the raw 45WS-WSR data were and horizontal radar cross sections on the 10 days gridded to a Cartesian coordinate system using the analyzed in this study. For the threshold-level events, 4 Python ARM Radar Toolkit (Py-ART) (Helmus and Collis multicell downburst events were selected before 10 2016). For the gridding process, a horizontal and other downburst events were randomly-selected. All 4 vertical grid resolution of 500 m was used along with a null events were selected on days during which constant 1 km radius of influence. A Cressman threshold-level events also occurred. weighting function (Cressman 1959) was used for data interpolation. Data were gridded out 100 km in the north-south and east-west directions and 17 km in the vertical direction, with the 45WS-WSR as the grid origin. 3. RESULTS AND DISCUSSION 3.1 Overview Four radar signatures common amongst threshold- level downburst-producing storm systems at CCAFS/KSC have been identified in this study: 1) Peak height of the 1 dB Zdr contour within a Zdr column; 2) Peak height of co-located values of 30 dBZ Zh and near-0 dB Zdr; 3) Peak Zh value within the storm system; 4) Gradient in Zdr below the 0 °C level within a DRC, including an increase in Zdr to a value of at least 3 dB. These signatures imply the importance of precipitation ice in threshold-level downbursts at CCAFS/KSC, as discussed below. The lead times offered by these signatures were relatively large in multicell events, due to the multiple updraft-downdraft cycles observed. 3.2 Signature #1 – Peak Zdr Column Height The first radar signature identified was the peak heights of Zdr columns within threshold-level downburst- producing thunderstorms. An example of this signature is presented in Fig. 2, where the 1 dB Zdr column can be seen approximately 16 to 20 km west of the radar extending to roughly 7.5 km above ground level (AGL), or about 3 km above the 0 °C level. The signature seen in Fig. 2 was present before a 35-knot downburst wind gust was recorded on Cape WINDS Tower 1007 about 48.5 minutes later at 2322:30 UTC. Since positive Zdr values are typical of oblate-shaped liquid hydrometeors and Zdr values around 0 dB are typical of spherical or ice hydrometeors (Herzegh and Jameson 1992), Zdr columns indicate regions where liquid hydrometeors are lofted above the environmental 0 °C level by the storm’s updraft (Tuttle et al. 1989). These lofted liquid hydrometeors will either freeze and subsequently melt as they descend back below the 0 °C level and/or evaporate which, along with mass loading, will Figure 2. Vertical cross sections taken in the east-west contribute to negative buoyancy within the storm and direction (i.e., x-z plane shown) 20.0 km north of the lead to formation and intensification of the downburst 45WS-WSR during the volume scan that ended around (Srivastava 1987). 2234 UTC on 19 June 2015. The top image shows radar In this study, it was observed that the 1 dB contour reflectivity factor in dBZ, the center image shows within a Zdr column extended at least 1 km above the differential reflectivity in dB, and the bottom image environmental 0 °C level in 12 of the 14 (85.71%) shows correlation coefficient in percentage form. In threshold-level events analyzed. This suggests that the each image, the black horizontal line marks the 0 °C lofting of liquid hydrometeors above the 0 °C level and level and the purple horizontal line marks the height of the subsequent freeze-melting, evaporation, and mass minimum equivalent potential temperature calculated loading processes are important to the formation of from the 0000 UTC KXMR sounding on 20 June 2015. threshold-level downbursts at CCAFS/KSC. The lead times offered by this signature ranged by the 45WS for events that meet their warning from 11.50 minutes to 78.50 minutes, with a mean lead Thrsehold-2. Therefore, this signature may offer the time of 40.67 minutes and a median lead time of 42.50 lead times desired by the 45WS for threshold-level minutes. Thus, the mean lead time offered by the 1 dB downbursts. However, a peak 1 dB Zdr column height of Zdr contour extending at least 1 km above the 0 °C level at least 1 km above the 0 °C level was also observed in is more than 10 minutes longer than desired by the 4 of the 4 (100%) null events analyzed in this study. 45WS for wind gusts that meet their warning Threshold- Therefore, while the 1 dB Zdr column height may offer 1. Furthermore, the maximum observed lead time for increased lead times for threshold-level downbursts, it this signature is nearly 20 minutes longer than desired may not serve as well for distinguishing storm cells that will produce threshold-level events from those that will co-located values of 30 dBZ Zh and near-0 dB Zdr can not. This signature could possibly lead to a high FAR for be seen from 3 to 10 km north of the 45WS-WSR 45WS convective wind warnings if used as the sole extending to 10 km AGL, or about 5 km above the 0 °C downburst prediction parameter, but may be very useful level. The signature seen in Fig. 3 was present before a if used in combination with other signatures. 42-knot downburst wind gust was recorded on Cape WINDS Tower 0421 about 50.5 minutes later at 2122:30 3.3 Signature #2 – Height of 30 dBZ Zh and 0 dB Zdr UTC. This signature implies the presence of a significant amount of precipitation ice aloft in the storm The second radar signature identified in this study cell. Past studies have shown that graupel and small was the peak height of approximately co-located values hailstones typically have a Zh ≥ 29 – 33 dBZ (Deierling of 30 dBZ Zh and near-0 dB Zdr. An example of this et al. 2008), so a Zh value of 30 dBZ was used in this signature can be seen in Fig. 3. In Fig. 3, the region of study as the lower boundary for precipitation ice. Zdr values within regions of precipitation ice are generally around 0 dB due to the spherical shape of the ice hydrometeors and/or the lower dielectric within the ice hydrometeors as a result of their lower bulk densities (Herzegh and Jameson 1992). Past studies have shown that the melting of ice hydrometeors may contribute more greatly to negative buoyancy within a downburst than evaporation of liquid hydrometeors, despite the latent heat of vaporization being roughly 8.5 times larger than latent heat of fusion, especially in regions where high relative humidity values enhance melting and suppress evaporation (Srivastava 1987). Therefore, the presence of a large quantity of precipitation ice within a storm cell may indicate the potential for negative buoyancy to be enhanced within the downburst via latent heat of melting and downward forcing from mass loading (Srivastava 1987). In this study, the co-located values of 30 dBZ Zh and near-0 dB Zdr extended at least 3 km above the 0 °C level in 13 of the 14 (92.86%) threshold-level events analyzed. This suggests that ice precipitation is important to the formation of threshold-level downbursts at CCAFS/KSC, as indicated by several past studies. The lead times offered by this signature ranged from 3.50 to 78.50 minutes, with a mean lead time of 40.88 minutes and a median lead time of 35.50 minutes. As with Signature #1, the mean lead time offered by co- located values of 30 dBZ Zh and near-0 dB Zdr was more than 10 minutes longer than desired by the 45WS for Threshold-1 events and the maximum lead time offered by this signature was nearly 20 minutes longer than desired for Threshold-2 events. However, co-located values of 30 dBZ Zh and near-0 dB Zdr extended at least 3 km above the 0 °C level in 4 of the 4 (100%) null events analyzed. Therefore, as with Signature #1, Signature #2 may be very useful for offering the increased lead times desired by the 45WS for threshold- level events, but may cause FAR to be relatively high if used alone. 3.4 Signature #3 – Peak Zh Value Figure 3. Vertical cross sections taken in the north- The third radar signature identified in this study was south direction (i.e., y-z plane shown) 28.5 km west of the peak value of Zh within the threshold-level the 45WS-WSR during the volume scan that ended downburst-producing storm system. An example of this around 2032 UTC on 12 September 2015. The variables signature can be seen in Fig. 4, where the peak Zh is and horizontal lines shown are the same as in Fig. 1, around 50 – 55 dBZ. The signature seen in Fig. 4 was with the environmental conditions calculated using the present before a 35-knot downburst wind gust was 0000 UTC KXMR sounding on 13 September 2015. recorded on Cape WINDS Tower 0001 about 24.5 minutes later at 1912:30 UTC. liquid hydrometeors can enhance negative buoyancy through latent heat absorbed by these processes (Srivastava 1987). For this study, only the peak Zh value was identified; the location of peak Zh relative to the 0 °C level will be examined in future work. A peak Zh of at least 50 dBZ was present in 13 of the 14 (92.86%) threshold-level events analyzed, which suggests the importance of a large quantity of hydrometeors in producing threshold-level downbursts at CCAFS/KSC. The lead times offered by this signature ranged from 11.50 to 78.50 minutes, with a mean lead time of 45.88 minutes and a median lead time of 48.50 minutes. The mean lead time offered by the peak Zh value within a storm system, as with the previous two signatures, is more than 10 minutes longer than desired by the 45WS for Threshold-1 events and the maximum lead time is nearly 20 minutes longer than desired for Threshold-2 events. However, a peak Zh value of at least 50 dBZ was observed in 3 of the 4 (75%) null events analyzed. This indicates that peak Zh of at least 50 dBZ in a storm system may offer longer lead times for threshold-level downbursts, but may lead to a high FAR if used as the sole predictor of threshold-level events. 3.5 Signature #4 – Zdr Gradient within the DRC The final radar signature identified in this study was the gradient in Zdr within the DRC of a downburst- producing storm cell. An example of this signature is provided in Fig. 5, where Zdr values can be seen increasing with decreasing height from about -1 dB to values greater than 3 dB within the DRC located about 41 – 44 km north of the 45WS-WSR around 2 – 3 km AGL. The signature seen in Fig. 5 was present before a 51-knot downburst wind gust was recorded on Cape WINDS Tower 0019 16.5 minutes later at 2132:30 UTC. Note that Tower 0019 is not shown in Fig. 1, but is located approximately halfway between Towers 0022 and 0015. This signature indicates the melting of small Figure 4. Vertical cross sections taken in the east-west precipitation ice hydrometeors (e.g., graupel) over a direction (i.e., x-z plane shown) 6.5 km north of the relatively shallow layer beneath the 0 °C level. Since Zdr 45WS-WSR during the volume scan that ended around is a measure of the reflectivity-weighted mean diameter 1848 UTC on 14 July 2015. The variables and of a given drop size distribution (DSD) (Jameson 1983), horizontal lines shown are the same as in Figs. 1 – 2, and smaller ice hydrometeors typically melt completely with the environmental conditions calculated using the before larger ice hydrometeors do, an increase in Zdr 1500 UTC KXMR sounding on 14 July 2015. from near-0 dB around the 0 °C level to positive values with decreasing height below the 0 °C level implies the This signature implies the presence of large-sized melting of smaller ice hydrometeors (Meischner et al. hydrometeors and/or a large concentration of 1991). Given that the latent heat absorbed by melting hydrometeors within the storm cell. A region of large Zh ice hydrometeors may contribute more to negative that develops above the 0 °C level implies the presence buoyancy within a downburst than the latent heat of a large quantity of precipitation ice in the cell, while a absorbed by evaporating liquid hydrometeors in regions region of large Zh that develops below the 0 °C level where relative humidity values are high (Srivastava implies the presence of a large quantity of liquid 1987), regions where Zdr increases sharply over a hydrometeors within the cell (Loconto 2006). The shallow layer below the 0 °C level may suggest the presence of a large quantity of hydrometeors (liquid melting of a large quantity of smaller ice hydrometeors, and/or ice) can enhance the negative buoyancy within a which can lead to significant downburst intensification downdraft through mass loading (Srivastava 1987). (Meischner et al. 1991). Past studies have also noted Additionally, melting from the large amount of ice that Zdr values increase from near-0 dB around the 0 °C hydrometeors and evaporation from the large amount of level to at least 3 dB below the 0 °C level within downburst-producing storms in warm humid climate minutes. As with the previous signatures, the mean lead regions (White 2015). time offered by this signature is more than 10 minutes longer than desired by the 45WS for Thrsehold-1 events and the maximum lead time was nearly 20 minutes longer than desired by the 45WS for Threshold-2 events. However, an increase in Zdr to at least 3 dB within 2.5 km below the 0 °C level was observed in the DRC of 4 of the 4 (100%) null events analyzed. Thus, the gradient in Zdr below the 0 °C level, including an increase in Zdr to at least 3 dB, may provide the lead times desired by the 45WS for threshold-level downburst events, but may also lead to a high FAR if used as the sole downburst prediction parameter. 4. SUMMARY AND FUTURE WORK The purpose of this study was to identify C-band dual-polarization radar signatures that can be used in real-time by 45WS forecasters to increase lead times and decrease FAR for their convective wind warnings and help distinguish which storms will produce a Threshold-1 peak wind gust (35+ knots) from those that will also produce a Threshold-2 peak wind gust (50+ knots). So far, four common radar signatures have been identified amongst threshold-level downburst-producing storm systems at CCAFS/KSC: 1) The 1 dB Zdr contour within a Zdr column extending at least 1 km above the 0 °C level; 2) Co-located values of 30 dBZ Zh and near-0 dB Zdr extending at least 3 km above the 0 °C level; 3) Peak Zh value of at least 50 dBZ within the storm system; 4) Zdr increasing to at least 3 dB within 2.5 km below the 0 °C level in the DRC. The mean lead times offered by each of these signatures ranged from 40 – 46 minutes, with median lead times between 35 – 49 minutes. The longer lead times were mainly due to the large number of multicell events within the storm sample, each of which contained several updraft-downdraft cycles and allowed these signatures to be studied well in advance of the threshold-level gust recorded on the Cape WINDS. These signatures are all related to precipitation ice, Figure 5. Vertical cross sections taken in the north- which suggests that the melting and mass loading of south direction (i.e., y-z plane shown) 16.0 km east of precipitation ice is very important in the formation of the 45WS-WSR during the volume scan that ended threshold-level downbursts at CCAFS/KSC. around 2116 UTC on 21 May 2015. The variables and This work is part of an ongoing research project. horizontal lines shown are the same as in Figs. 1 – 3, While the radar signatures identified in this study may with the environmental conditions calculated using the offer the lead times desired by the 45WS, future work in 0000 UTC KXMR sounding on 22 May 2015. this project will include identifying signatures that are more exclusive to threshold-level events and less In this study, it was observed that Zdr values common in null events in an effort to decrease FAR for increased from around 0 dB near the 0 °C level to at 45WS convective wind warnings. Signatures will also be least 3 dB within 2.5 km below the 0 °C level in a DRC examined that may help distinguish Threshold-1 events in 13 of the 14 (92.86%) threshold-level events from Threshold-2 events. Given the relatively small analyzed. This indicates the importance of the melting of sample size of 14 threshold-level events and 4 null smaller ice hydrometeors to the intensification of events used in this study, more cases will be added to threshold-level downbursts at CCAFS/KSC. each of these categories in the near future to further The lead times offered by this signature ranged examine the occurrence of these four radar signatures from 1.50 minutes to 78.50 minutes, with a mean lead and any other new signatures that are identified. time of 40.42 minutes and a median lead time of 41.50 Other future work will include examining the Loconto, A. N., 2006: Improvements of Warm-Season location of peak Zh relative to the 0 °C level in threshold- Convective Wind Forecasts at the Kennedy Space level events. Through not discussed herein, several Center and Cape Canaveral Air Force Station. M.S. events in this study occurred within the presence of sea Thesis, Dept. of Chemical, Earth, Atmospheric, and breeze fronts, gust fronts, and/or storm mergers. The Physical Sciences, Plymouth State University, presence of these features may also be useful in Plymouth, NH, 79 pp. forecasting threshold-level downbursts at CCAFS/KSC. Furthermore, environmental data will be examined in Meischner, P. F., Bringi, V. N., Heimann, D., and Holler, more detail to better understand how certain conditions H., 1991: A Squall Line in Southern Germany: (e.g., relative humidity profile, temperature lapse rate, θe Kinematics and Precipitation Formation as profile) can be used to meet the three primary goals of Deduced by Advanced Polarimetric and Doppler this study. The results of this research project will Radar Measurements. Mon. Wea. Rev., 119, 678- eventually be integrated into a new algorithm that can 701. be used by 45WS forecasters to aid in the forecasting of downbursts at CCAFS/KSC. Roeder, W. P., McNamara, T. M., Boyd, B. F., and Merceret, F. J., 2009: The New Weather Radar for 5. ACKNOWLEDGEMENTS America’s Space Program in Florida: An Overview. 34th Conference on Radar Meteorology, This work has been supported by *INSERT Williamsburg, VA, Amer. Meteor. Soc., 10B.6. FUNDING INFO HERE*. We would like to thank Mr. [Available online at Jeffrey Zautner of the 14th Weather Squadron for http://kscwxarchive.ksc.nasa.gov/Publications]. providing the Cape WINDS and KXMR data. The first author would like to thank: Dr. Lawrence Carey of the -----, Huddleston, L. L., Bauman III, W. H., and Doser, K. University of Alabama in Huntsville for his support and B., 2014: Weather research requirements to guidance throughout this study, Mr. William Roeder and improve space launch from Cape Canaveral Air Mr. Todd McNamara of the 14WS for providing support Force Station and NASA Kennedy Space Center. for this study and the 45WS-WSR data, Dr. Richard 2014 Space Traffic Management Conference, Blakeslee of NASA Marshall Space Flight Center for Daytona Beach, FL, Embry-Riddle Aeronautical providing funding for this study, and members of the University, Paper 14, [Available online at Severe Weather Institute and Radar & Lightning http://commons.erau.edu/cgi/viewcontent.cgi?article Laboratories for their help throughout this study. =1027&context=stm]. 6. REFERENCES Scharfenberg, K. A., 2003: Polarimetric Radar Signatures in Microburst-Producing Thunderstorms. Cressman, G. P., 1959: An operational objective 31st Int. Conf. on Radar Meteorology, Seattle, WA, analysis. Mon. Wea. Rev., 87, 367-374. Amer. Meteor. Soc., 8B.4. [Available online at https://ams.confex.com/ams/32BC31R5C/techprogr Deierling, W., Petersen, W. A., Latham, J., Ellis, S., and am/paper_64413.htm]. Christian, H. J., 2008: The relationship between lightning activity and ice fluxes in thunderstorms. J. Smith, T. M., Elmore, K. L., and Dulin, S. A., 2004: A Geophys. Res., 113, D15210. Damaging Downburst Prediction and Detection Algorithm for the WSR-88D. Wea. Forecasting, 19, Gibson, M., 2005: GR2 Analyst. Gibson Ridge Software, 240 – 250. LLC. Srivastava, R. C., 1987: A Model of Intense Downdrafts Helmus, J. J., and Collis, S. M., 2016: The Python ARM Driven by the Melting and Evaporation of Radar Toolkit (Py-ART), a Library for Working with Precipitation. J. Atmos. Sci., 44, 1752 – 1773. Weather Radar Data in the Python Programming Language. Journal of Open Research Software, 4, Tuttle, J. D., Bringi, V. N., Orville, H. D., and Kopp, F. J., e25, 1 – 6. 1989: Multiparameter Radar Study of a Microburst: Comparison with Model Results. J. Atmos. Sci., 46, Herzegh, P. H., and Jameson, A. R., 1992: Observing 601 – 620. Precipitation through Dual-Polarization Radar Measurements. Bull. Amer. Meteor. Soc., 73, 1365 Wakimoto, R. M., and Bringi, V. N., 1988: Dual- – 1374. Polarization Observations of Microbursts Associated with Intense Convection: The 20 July Jameson, A. R., 1983: Microphysical Interpretation of Storm during the MIST Project. Mon. Wea. Rev., Multi-Parameter Radar Measurements in Rain. Part 116, 1521 – 1539. I: Interpretation of Polarization Measurements and Estimation of Raindrop Shapes. J. Atmos. Sci., 40, White, P. W., 2015: An Exploratory Study in Nowcasting 1792 – 1802. Convective Winds using C-band Dual Polarimetric Radar. M.S. Thesis, Dept. of Atmospheric Science, The University of Alabama in Huntsville, Huntsville, AL, 104 pp