1.4 GLOBAL STUDIES OF TROPICAL CYCLONES USING THE WORLD WIDE LIGHTNING LOCATION NETWORK

Natalia N. Solorzano1*, Jeremy N. Thomas2,3, and Robert H. Holzworth2 1Digipen Institute of Technology, Redmond, WA, USA 2University of Washington, Seattle, WA, USA 3USGS, Denver, CO, USA

1. INTRODUCTION 1 2. THE WORLD WIDE LIGHTNING LOCATION NETWORK The lightning activity generated by tropical cyclones is not well understood, mainly because these storms occur The World Wide Lightning Location Network (WWLLN; over the ocean away from land-based regional lightning http://www.wwlln.net/) allows us to investigate lightning networks. Recent studies have used satellite-based activity driven by tropical cyclones that occur anywhere lightning detection (e.g. OTD, LIS, and FORTE) and on earth. Hence, this work is unique from any past extended regional networks (e.g. LLDN) to investigate studies of tropical cyclones that were limited to regions lightning activity in tropical cyclones well before they within a few hundred kilometers of ground-based reach . However, these satellites can only detect sensors or utilized a few minutes of satellite-based data lightning from a particular storm for a few minutes each each day. day and these extended networks only cover a limited global region. The WWLLN provides real time lightning locations globally by measuring the very low frequency (VLF) In this paper, we use the World Wide Lightning Location Network (WWLLN), the only real-time network that radiation (3-30 kHz) emanating from lightning covers the entire globe, to analyze the change in discharges. For a lightning stroke to be accurately lightning activity during the evolution of tropical detected with error analysis, the VLF radiation from the cyclones. We present case studies of tropical cyclones stroke must be detected at a minimum of 5 of the in different global regions to investigate whether network’s 26 receivers around the world. Each receiving lightning activity can be used as a proxy for convection station consists of a whip antenna to measure VLF evolution and organization. For each case study, we electric field, a GPS antenna for accurate timing, discuss the spatial and temporal lightning activity, in preamplifying electronics, and an internet-connected regions such as in the eyewall and rainband. These processing computer. Each receiver locally processes a case studies will be compared to the previous work of stroke’s waveform and sends the time of group arrival Molinari et al. (1999), which suggests that eyewall (TOGA) to the central processing station for location lightning intensifies before eyewall replacement cycles (Dowden et al., 2002). In this manner, WWLLN during Atlantic basin hurricanes. Hence, we examine provides continuous lightning detection coverage of the whether lightning activity can remotely identify the entire globe. eyewall replacement process. These eyewall cycles are often covered by thick cirrus overcast and thus can The location accuracy and efficiency of WWLLN have typically only be observed using in situ aircraft been estimated for certain regions of the globe by measurements. In this work, we use WWLLN to comparison to regional, ground-based lightning investigate the lightning activity during the entire life detection systems (Lay et al. 2004; Rodger et al. 2004; span of Atlantic basin hurricanes and Pacific basin Jacobson et al. 2006). Rodger et al. (2004) completed a . comparison of WWLLN data in Australia to the local Additionally, to investigate how tropical cyclones Australian lightning location network, Kattron, and found sometimes intensify or weaken just before making a detection efficiency of ~26% of cloud-to-ground (CG) landfall, we study the lightning activity as the storms strokes in Australia and ~10% of intracloud (IC) strokes, approach and hit land. Finally, we discuss whether with a location error of 4.2 +/- 2.7 km. monitoring lightning activity during tropical cyclones can lead to better forecasting and nowcasting of storm 3. CASE STUDIES evolution, especially where aircrafts measurements are In this work, we present results concerning the spatial not typically feasible. and temporal distributions for two Atlantic basin hurricanes, and three western Pacific typhoons. To the authors’ knowledge, these are the first results concerning the lightning activity in which the entire life span of the typhoons is observed.

*Corresponding author address: Natalia N. Solorzano, Physics Dept., Digipen Institute of Technology, Redmond, WA 98052; e-mail: [email protected]

1 In order to test if our results obtained from WWLLN data are representative of the lightning distribution, we plot the lightning spatial and temporal distributions for Hurricane Rita (Figure 1) and compared to previous studies. Rita was a category 5 hurricane, the fourth most intense Atlantic cyclone ever recorded in the Gulf of Mexico.

Figure 1 (top) shows WWLLN data for September 21, 2005 when Rita rapidly intensified from a category 2 to 5 storm. The WWLLN data show the radial pattern observed by Molinari et al. (1999), for Atlantic basin hurricanes. There are three distinct regions: a weak density maximum for the eyewall (region within ~40 km of the center), a distinct area of minimum activity at approximately 80 – 200 km from the eyewall, and the main, broader maximum on the rainband region, outside the 200 km radius.

The spatial distribution of lightning activity for Rita’s eyewall is shown in Figure 1 (middle). The lightning frequency was integrated over the period of 14:00 - 15:00 UT of September 21, 2005, which was during very rapid strengthening. It is possible to observe the eyewall activity. According to Molinari et al. (1994 and 1999), the eyewall lightning is episodic, and activity outbreaks may accompany the eyewall cycles. Our results are in good agreement with the observations from the LASA lightning network reported by Shao et al. (2005)

Figure 1 (bottom) shows the time histogram of Hurricane Rita eyewall lightning (here defined as within 100 km of the storm center) obtained by WWLLN for September 18-27, 2005, along with and minimum pressure1. We observed that at, or prior to, most intensity changes, indicated by the change in the slope on the maximum sustained wind (or the minimum pressure), there is a lightning outbreak. These peak lightning activities accompanying intensity changes for Rita were also reported by Squires (2006) using the LLDN network. Figure 1 indicates a large increase in eyewall lightning preceding the eyewall replacement that occurred on September 22 as reported by Houze et al. (2007). Additionally, just prior to and during landfall on September 24 there was a weak eyewall lightning outbreak.

The results presented in Figure 1 show that the data from WWLLN are in good agreement with previous works, indicating that, although WWLLN is most sensitive to large peak current strokes (above 30 kA), the data set provides the information we need to study Figure 1: (top) WWLLN lightning in storm centered the lightning activity for tropical cyclones. coordinates during Hurricane Rita for UT day Sept. 21, 2005. (middle) A zoom of the eyewall lightning for 14:00 – 15:00 UT Sept. 21. (bottom) Time histogram of eyewall lightning (within 100 km of the storm center) for Sept. 18- 27, 2005 along with maximum sustained wind and minimum pressure.

1 wind and pressure data obtained from http://www.solar.ifa.hawaii.edu/Tropical/tropical.html

2 Figure 2: (top) WWLLN lightning in storm centered Figure 3: (top) WWLLN lightning in storm centered coordinates during Hurricane Katrina for UT day Aug coordinates during for Nov. 30 28, 2005. (bottom) Time histogram of Hurricane 2006. (bottom) Time histogram of Katrina eyewall lightning (within 100 km of the storm eyewall lightning (within 100 km of the storm center) center) observed by WWLLN for Aug 24-31, 2005 observed by WWLLN for Nov. 26 – Dec 6, 2006 along with maximum sustained wind and minimum along with maximum sustained wind. pressure. Next, we discuss three western Pacific super typhoons Figure 2 (top) shows the lightning density obtained by (equivalent to category 4-5 hurricanes): Durian, WWLLN for Hurricane Katrina at the peak intensity day Chanchu, and Yagi. (UT day Aug 28, 2005). Katrina was a category 5 hurricane, and also one of the five most intense Figure 3 (top) shows WWLLN lightning for Durian, which hurricanes in the history of the United States. Again, made landfall in the and , during the same radial pattern with 3 distinct regions described on November 30, 2006. It is by Molinari et al. (1999) can be observed. possible to observe the 2 maxima and the minimum. Figure 3 (bottom) shows the lightning frequency for Figure 2 (bottom) shows the frequency for Hurricane Durian, along with maximum sustained wind. Peaks in Katrina eyewall lightning observed by WWLLN for Aug activity are observed before or during most major 24-31, 2005 along with maximum sustained wind and intensity changes. The largest peak occurs prior to the minimum pressure. The first landfall, as indicated, first landfall. occurred on August 25, in Florida, when Katrina was a moderate category 1 storm. Interestingly, after this Figure 4 (top) shows the spatial lightning distribution for event, the storm intensified, and this process was Chanchu during rapid intensification (May 14, 2006). accompanied by a peak in lightning activity. A second Again, three regions of distinct density can be identified. and a third landfall occurred on August 29, in Louisiana, At the bottom of Figure 4, the lightning spatial when Katrina was a category 3 hurricane. After the distribution is depicted. Chanchu struck the Philippines , the storm weakened, but no change in the twice as a typhoon, upgraded to a super typhoon while lightning activity was observed. in the ocean, and hit after weakening. Most of the

3 Figure 4: (top) WWLLN lightning in storm centered Figure 5: (top) WWLLN lightning in storm centered coordinates during for May 14, coordinates during Typhoon Yagi for Sept. 21, 2006. 2006. (bottom) Time histogram of Typhoon Chanchu (bottom) Time histogram of Typhoon Yagi eyewall eyewall lightning (within 100 km of the storm center) lightning (within 100 km of the storm center) observed observed by WWLLN for May 9-18, 2006 along with by WWLLN for Sept. 17 - 24, 2006 along with maximum sustained wind. maximum sustained wind. studied, peaks on lightning time histograms occur prior intensity changes are accompanied by peaks in the or during major intensity changes, indicating that lightning frequency. lightning activity might be used as a proxy for convection intensity in strong tropical cyclones. Peaks in Figure 5 (top) shows WWLLN data for Typhoon Yagi, a lightning frequency can also be detected before super typhoon, on the peak intensity day (September landfalls. 21, 2006). A similar pattern for lightning density observed for the previous storms is seen here. Figure 5 Since Rita and Katrina were Atlantic basin storms that (bottom) contains the lightning time histogram plotted made landfall in the US, they are well studied and with wind intensity for Yagi. Peaks in lightning activity numerous observations are available, including in situ can be observed during most of the intensity changes. aircraft measurements, ground-based and satellite- based radar, and satellite imagery. We plan to compare 4. CONCLUSIONS AND FUTURE WORK the WWLLN lightning observed during Rita and Katrina to establish lightning activity as a proxy for various The WWLLN data set was used to study the lightning stages in tropical cyclonic evolution. To develop a more temporal and spatial distributions for two hurricanes, robust, statistically significant relation between lightning Rita and Katrina, and three typhoons, Durian, Chanchu activity and storm evolution, we will examine all of the and Yagi. The results obtained for Rita were compared tropical cyclones that occurred globally since 2004, to previous works, and similar results were obtained. when WWLLN became fully operational. We also plan Western Pacific tropical cyclones were studied for the investigate the eyewall lightning location with respect to first time in terms of lightning. For all storms observed, the vertical wind shear vector to compare with the we could identify three regions of distinct spatial observations of Molinari et al., 2004, 2006. We also behavior for lightning density. Also for all the cases

4 intend to compare the results obtained for western Squires, K. A., 2006: The morphology of eyewall cloud Pacific storms to the available weather data. to ground lightning in two category five hurricanes, MS Thesis, U. of Hawaii.

5. ACKNOWLEDGMENTS

We thank WWLLN station hosts throughout the globe.

6. REFERENCES

Dowden, R.L., J.B. Brundell, and C.J. Rodger, 2002: VLF lightning location by time of group arrival (TOGA) at multiple sites, Journal of Atmospheric and Solar- Terrestrial Physics, 64 (7), 817-30.

Houze, R.A. Jr., S. S. Chen, B. F. Smull, W. Lee, and M. M. Bell, 2007: Hurricane intensity and eyewall replacement, Science, 315, 1235-1239.

Jacobson, A.R., R.H. Holzworth, J. Harlin, R.L. Dowden, and E. H. Lay, 2006: Performance assessment of the World Wide Lightning Location Network (WWLLN), using the Los Alamos Sferic Array (LASA) array as ground-truth, Journal of Atmospheric and Oceanic Technology, 23, 1082-92.

Lay, E.H., R.H. Holzworth, C.J. Rodger, J.N. Thomas, O. Pinto, Jr., and R.L. Dowden, 2004: WWLL Global Lightning Detection System: Regional Validation Study in Brazil, Geophys. Res. Lett., 31, L03102, doi: 10.1029/2003GL018882.

Molinari, J., P.K. Moore, V.P. Idone, R.W. Henderson, and A.B. Saljoughy, 1994: Cloud-to-ground lightning in Hurricane Andrew, J. of Geophys. Res., 99 (D8), 16665- 16676.

Molinari, J., P.K. Moore, and V.P. Idone, 1999: Convective structure of hurricanes as revealed by lightning locations, Monthly Weather Review, 127 (4): 520-534.

Molinari, J., D. Vollaro, and K.L. Corbosiero, 2004: formation in a sheared environment: A case study. Journal of the Atmospheric Sciences, 61, 2493-2509.

Molinari, J., N.W.S. Demetriades, R.L. Holle, and D. Vollaro, 2006: Application of long-range lightning data to hurricane formation and intensification. Second Conf. on Meteor. Applic. Of Lightning Data, Atlanta, Georgia.

Rodger, C. J., J. B. Brundell, R. L. Dowden, and N. R. Thomson, 2004: Location accuracy of long distance VLF lightning location network. Ann. Geophys., 22, 747-758.

Shao, X.M., et al., 2005: Katrina and Rita were lit up with lightning, Eos, 86 (42), 398.

5