Abstract

Lightning data (cloud-to-ground plus intracloud) obtained from the Los Alamos Sferic Array (LASA) for 2005’s Hurricanes Rita and Katrina were analyzed to provide a first insight into the three-dimensional electrical activity of rapidly intensifying hurricanes. This information is crucial for modelers aiming at better forecasting hurricane intensity, because it is inherently related to key structural aspects of the storm often misrepresented in numerical models. Analysis of the intracloud narrow bipolar events (NBEs) for Rita revealed a general increase in discharge heights during the period of rapid intensification. The results also showed that for the case of Rita, NBEs were useful in tracking and mapping the evolution of individual strong convective elements embedded in the eyewall during rapid intensification. Those results are particularly revealing, and suggest that the general increase in height of the intracloud lightning is an aggregate consequence of numerous short-lived convective events rotating rapidly around the eyewall of Rita. A similar rise in discharge heights during periods of intensification was also observed for Katrina. However, the NBE lightning data show that for Katrina, the eyewall convection persisted for several hours at a fixed location instead of rotating cyclonically along the eyewall. This highlights the idea that NBE lightning data can also be used to identify different convective regimes attributed to possibly different internal or external forcing mechanism(s).

1. Introduction

Hurricanes are one of the most destructive natural phenomena and are regularly related to the loss of life as well as extensive property damage. In recent decades, significant progress has been made toward forecasting their track via a better understanding of the overall storm dynamics, improved model simulations, and by developing more advanced measurement technologies. However, relatively little progress has been made in forecasting hurricane intensity (e.g., Marks and Shay 1998). This is because observational (e.g., Simpson et al. 1998; Rodgers et al. 1998; Reasor et al. 2009; Guimond et al. 2010) and modeling (Hendricks et al. 2004; Montgomery et al. 2006; Fierro and Reisner 2011) studies have suggested that the physical processes leading to the intensification of hurricanes are believed to not only rely on large-scale dynamics, but also on small-scale, hard-to-observe, transient convective bursts (or vortical hot towers) occurring within the eyewall. Most of the current observing platforms, such as radar from hurricane hunter aircraft, do not provide temporally continuous coverage of these smaller-scale individual convective bursts. These lack of observations could put a limit on predicting rapid intensification (RI; Kaplan and DeMaria 2003; Kaplan et al. 2010) of hurricanes.

Recently, studies have shown that lightning bursts often occur prior to intensification. Price et al. (2009) analyzed the total lightning flash rate (cloud-to-ground plus intracloud) of 56 hurricanes around the globe using the World Wide Lightning Location Network (WWLLN), which is primarily sensitive to the most intense lightning strokes (Jacobson et al. 2006; Lay et al. 2007). Their study revealed a strong correlation between maximum sustained winds and total flash rate. While it is possible that not all intensifying hurricanes experience eyewall lightning outbreaks (Kelley et al. 2004), observations of hurricane eyewall lightning activity and its possible relationship to hurricane intensification have been reported (e.g., Molinari et al. 1994, 1999). Similar findings have been observed through more focused works on Hurricanes Katrina and Rita (2005) using the Los Alamos Sferic Array (LASA; Shao et al. 2005), WWLLN (Thomas et al. 2010), and the National Lightning Detection Network (NLDN; Squires and Businger 2008). A more comprehensive study by DeMaria and DeMaria (2009) with WWLLN suggested that, on the other hand, lightning outbreaks at the outer radii, and not in the eyewall, might be a better surrogate for hurricane intensification.

This work builds upon the above studies by showing that eyewall lightning, cloud-to-ground (CG) and intracloud (IC) flashes together, can be used to provide insights into the key structural aspects of the eyewall during intensification, such as size, vertical tilt, and convective state. A major benefit of the LASA array, which is described in further detail in the following section, is that it is indiscriminant when recording discharge waveforms from either CG or IC flashes (i.e., there is no type-specific bias in the lightning it locates other than any range–detection efficiency effects from the sources themselves). This advantage becomes evident when considering that the rate of both IC and CG flashes shows overall good correlation with convective strength, at least in continental storms (Reap and MacGorman 1989; Wiens et al. 2005). Hence, we expect that if the convective updraft is vigorous enough within the eyewall of a hurricane, a copious amount of CG and IC flashes would be produced, and some of the intense ICs could be detected by the LASA system. With the recorded waveforms, the height of a special type of intense IC lightning discharges, the narrow bipolar events (NBEs), can be determined (Smith et al. 1999).

NBE flashes are a distinct class of IC lightning discharges, which were first detected and reported by Le Vine (1980) and Willett et al. (1989). The signatures of their very low-frequency and low-frequency (VLF–LF) radiation signals are quite unique and distinguishable from other types of lightning, both IC and CG flashes. They are distinguished by their relatively narrow (∼10 μs) pulse widths, extremely large amplitudes (∼10 V m−1 electric field change at 100 km), and their temporal isolation for other lightning activity (often isolated in time by many hundreds of μs or more). The discharges are inferred to be quite small in physical dimension, with length scales estimated to be not more than a few kilometers (Smith et al. 1999; Eack 2004; Hamlin et al. 2007). What makes NBEs crucial for this study is that although very infrequent (comprising much less than one percent of all IC lightning) their signal amplitude is quite large, making them detectable from much further ranges than other forms of IC lightning, with detection efficiencies similar to that of first return strokes because of their similar VLF–LF radiation amplitudes (Smith et al. 1999; Eack 2004).

It is possible to reveal the detailed IC structure (and hence the storm electrification structure) inside the cloud if the majority of the small-amplitude IC signals are detected and geolocated. Unfortunately, the storms of interest in this study were many hundreds of kilometers away from the LASA sensors, and the majority of the LASA-detected IC discharges were NBEs. Nevertheless, NBEs were often found to be associated with the beginning of IC flashes (Rison et al. 1999), and could be considered as indicators for the electrically active regions in the storm.

Previous observations for continental storms found that updraft intensification was often accompanied by lofting of the charge layers, and hence IC flash heights (MacGorman et al. 1989). Moreover, NBE activity within continental thunderstorms was shown to be a good proxy for storm severity (Jacobson and Heavner 2005). Therefore, the LASA height observations for the NBEs can be used to probe the height of the active discharge regions, and therefore might be used as an indicator of convection strength.

This work will exemplify the possibilities of using IC flash data to remotely observe important structural aspects of intense hurricanes. The ability to use IC lightning data to track intense convective bursts or events (CEs) within a hurricane could provide both forecasters and the modeling community an additional source of data that could be used to assist local authorities in making informed decisions regarding, for instance, whether or not coastal evacuations should take place. It is preferred herein to refer to the bursts to as “convective events” because to the authors’ knowledge, no observations of vortical thunderclouds within Rita or Katrina have been published to date.

2. Data and instrumentation

LASA is a scientific research system that consists of eight VLF to LF sensors in Florida and six sensors over the Great Plains (Shao et al. 2006). Unlike the other lightning networks, LASA records and stores all triggered lightning signals, including the raw electric field amplitude as a function of time waveforms, which can be used later in postprocessing for a more detailed analysis of the events. During the time frame of Rita’s and Katrina’s RI periods, when they were near the middle of the Gulf of Mexico, three Florida sensors and all Great Plains sensors were operational and were used for the lightning observations presented herein. With known sensitivity and trigger threshold for each individual station, the detection sensitivity for the combined array was estimated and is shown in Fig. 1. As Katrina and Rita underwent RI, which was within about 100 km of 25°N, 86°W, LASA is expected to detect and locate any type of lightning strokes that produce a peak current greater than ∼25 kA, which corresponds to a ∼50% detection efficiency for the first negative and positive return strokes (Uman 1987). Peak current for normal IC flashes are difficult to estimate, but their associated field changes are regularly much smaller than those of return strokes, and the detection efficiency at this range is estimated to be ∼1%. However, the detection efficiency for NBEs is expected to still be at the same ∼50% level as first return strokes because their peak field change (and current) is comparable to that of return strokes (Shao et al. 2006). Smith et al. (1999) estimated the average peak current for compact intracloud discharges (CIDs; what NBEs were originally called) at 28 kA. Eack (2004) concluded an average of ∼20 kA. These values are right in line with the typical −CG return stroke current.

Fig. 1.

Map showing the minimum return-stroke-equivalent peak current (in kA) for a lightning event, which would be detected simultaneously by four or more LASA stations (denoted by the solid black dots). The solid contours are for daytime detection and the dashed for nighttime detection, after considering the effects of groundwave propagation and the day and nighttime ionospheric reflections (Shao and Jacobson 2009).

Fig. 1.

Map showing the minimum return-stroke-equivalent peak current (in kA) for a lightning event, which would be detected simultaneously by four or more LASA stations (denoted by the solid black dots). The solid contours are for daytime detection and the dashed for nighttime detection, after considering the effects of groundwave propagation and the day and nighttime ionospheric reflections (Shao and Jacobson 2009).

Lightning geolocation was performed via the standard time-of-arrival technique where accuracy relies on the position of the storm relative to the station location baseline geometry (e.g., Shao et al. 2006). In this work, Rita’s position during the period of intensification (25°N, 86°W) was almost ideal in that regard. Katrina, on the other hand, was located too far to the east at earlier times, almost directly due south from the Florida stations, resulting in relatively large errors in discharge location (on the order of 50 km or more) along a line with endpoints at the storm center and at the Florida stations. However, the geolocation error in a direction perpendicular to that line (nearly east–west) remains on the order of 10 km or less at that time for Katrina.

As discussed later, NBEs are the events that provide unique height information for the eyewall lightning activity. The discharge height calculation for an NBE is performed after the primary event location determination. To determine the location, three or more stations are needed by using the time-of-arrival technique. Once the location is determined, the source height can be estimated by examining the NBE time waveform from a single station measurement (Smith et al. 2004). For a narrow pulse (like those from NBEs), LASA observations can be used to identify the direct groundwave signal that propagates along the earth’s surface from the source to the receiver, and a pair of delayed ionospheric pulses that are reflected from the ionosphere’s D-layer (60–90-km altitude). The first ionospheric pulse propagates from the source upward and then bounces off of the ionosphere back toward the sensor, and the second pulse travels downward first to the ground, bounces off of the ground to the ionosphere, and then travels back downward to the sensor. With the known horizontal distance from the source to the sensor and the time separations among the three pulses, the NBE source height and the ionospheric reflection height can be independently determined with a single-station measurement.

The accuracy of the NBE height is a function of the NBE-to-sensor distance, which determines the incident angles for the ionospheric reflections and therefore the extent of ionospheric dispersion of the original signal (Shao and Jacobson 2009). More dispersion (longer distance) will stretch the reflected pulses wider in time, and will affect the time-tag accuracy of the pulses, and therefore the accuracy of the height determination. At the range of ∼500 km to the nearest Florida sensor, the systematic bias for the source height is estimated to be within ±1 km, and the relative error from one NBE to the next that originated from the same region to be within ±0.5 km (Smith et al. 2004). As a result of the favorable triangular geometry formed by Rita, the Florida sensors, and the Great Plains sensors, the location accuracy for the eyewall lightning is estimated to be roughly 10 km (Shao et al. 2006).

In this paper, LASA lightning data are compared against two datasets of hurricane tracks and pressure traces. This work makes use of the 6-hourly standard National Hurricane Center (NHC) best-track data (Landsea 1993; Landsea et al. 2004) and the 3-hourly data from the NHC public advisories (which are both publicly available online at http://www.nhc.noaa.gov/). It is important to note that the best-track database takes into account systematic and random errors pointed out by Neumann (1994) and most importantly corrected biases introduced by different analysis techniques and improved knowledge on tropical cyclones by the community (Landsea 1993). Although the exact magnitude of the track and intensity errors is not known, the above likely explains why for both Katrina and Rita, notable differences are seen in the minimum surface pressures between the two aforementioned datasets, which in turn influences the interpretations that are made on the relationships between observed lightning and storm intensity changes. While this might be a coincidence, the analysis of the LASA lightning revealed that the pressure trace from the public advisories is more consistent with the timing of the IC outbursts for the two storms presented herein, and also for another hurricane not shown in this paper (Hurricane Charley in 2004). For this reason, both pressure trace datasets were included in order to highlight this noteworthy difference without, however, favoring one dataset over the other.

The in situ Doppler radar data shown in this work for Hurricane Rita were gathered during research missions flown by the NOAA aircraft during 1900–1920 UTC 21 September 2005. The NOAA research aircraft is equipped with three radars, consisting of two 5.5-cm C-bands on the nose and lower fuselage that scan in the horizontal and one 3.2-cm X-band radar on the tail that scans in the vertical. Details behind the characteristics of the radars are provided in Table 5 of Jorgensen (1984). [The radar data and imageries are publicly available at the Hurricane Research Division (HRD) Web site at http://www.aoml.noaa.gov/hrd/.]

For the analysis, only the lightning data obtained within a 100-km radius from the NHC best-track estimated storm center were analyzed. This is because the key in the intensification of hurricanes is believed to lie in the internal core dynamics rather than the adjacent rainbands, which are also believed to play a role, although still considered controversial (e.g., Powell 1990; Wang 2009).

3. Hurricane Rita results

The first new result presented here is the detection of different types of lightning discharges within the eyewall of Hurricane Rita. Note that because the great majority (>95%) of detected ICs are NBEs due to the low-detection efficiency of ICs, the NBE acronym is used in the remainder of the text. An eyewall lightning outburst is defined as a flash rate equal or exceeding the average hourly flash rate by a factor of 2 or more.

Figure 2a compares the time evolution of LASA-detected NBE (labeled IC) and CG lightning discharges in the eyewall region with the minimum surface pressures for the time period starting at 0000 UTC 21 September 2005 and ending at 0300 UTC 22 September 2005, time during which Rita underwent RI. The start of the RI period, based on the minimum surface pressure trace estimate from the NHC public advisories, is at about 1500 UTC 21 September 2005 (Fig. 2a). Three noticeable lightning bursts can be seen: the first flare-up, dominated by CGs, started near 1400 UTC and peaked near 1600 UTC; a second outburst, dominated by NBEs, peaked near 2000 UTC; and the third burst, dominated by CGs, peaked at 0200 UTC 22 September 2005. Consistent with previous studies of this storm, the onset of the first CG lightning burst preceded the abrupt pressure drop by 3–4 h, suggesting that vigorous convective activity in the eyewall region was occurring prior to RI. We note also that the LASA CG time series is in good agreement with the one produced by NLDN (Squires and Businger 2008) and WWLLN (Solorzano et al. 2008; Thomas et al. 2010).

Fig. 2.

(a) Time series of intracloud (IC: light gray) and cloud-to-ground (CG: black) lightning hourly flash rate for Hurricane Rita overlaid with a scatterplot of the NHC best-track (black dots) and NHC public advisories estimated minimum central pressure (triangles). The SLP acronym on the right axis of (a) stands for sea level pressure (hPa). (b)–(g) Maps of 2-h accumulated IC (red) and CG (blue) lightning flashes for Rita overlain by the NHC best track in the thick black line. The UTC times shown define the period of analysis of the lightning.

Fig. 2.

(a) Time series of intracloud (IC: light gray) and cloud-to-ground (CG: black) lightning hourly flash rate for Hurricane Rita overlaid with a scatterplot of the NHC best-track (black dots) and NHC public advisories estimated minimum central pressure (triangles). The SLP acronym on the right axis of (a) stands for sea level pressure (hPa). (b)–(g) Maps of 2-h accumulated IC (red) and CG (blue) lightning flashes for Rita overlain by the NHC best track in the thick black line. The UTC times shown define the period of analysis of the lightning.

Figures 2b–g show the spatial evolution of the CG and NBE (again, labeled IC) events within the eyewall. During the intensification stage (1400–2000 UTC, Figs. 2b–d), lightning occurred in a nearly symmetric pattern around the eyewall (e.g., Shao et al. 2005; Solorzano et al. 2008; Thomas et al. 2010) similar to Hurricane Andrew (1992; Williams 1995). As the storm reached maximum intensity (2200–0000 UTC, Fig. 2e), and before the onset of the eyewall replacement cycle (0002 UTC 22 September 2005, Knabb et al. 2005), the symmetric pattern of the lightning was not as well defined (Fig. 2f). Within a short period of time (0200–0400 UTC) a more well-defined eyewall was mapped out by both NBE and CG lightning flashes (Fig. 2g). Although the dominant lightning type changed from CG to IC during the intensification of Rita (e.g., Figs. 2c,d), the respective polarity of NBE and CG flashes did not. The vast majority (>98%) of the CGs in Rita’s eyewall moved negative charge down to the surface (−CG), while all NBEs moved negative charge upward (+NBE, not shown). This simultaneous presence of −CGs at lower levels and +NBEs aloft indicates the existence of a normal tripolar gross charge structure in the convective cells of the eyewall, similar to that of a regular airmass thunderstorm (e.g., Williams 1989). A recent modeling study by Fierro et al. (2007) of electrification within an idealized hurricane also hypothesized that for hurricane eyewalls largely dominated by −CG flashes, the modeled gross eyewall charge structure should resemble that of a normal tripole.

The second new result is the three-dimensional view of the eyewall lightning activity. Starting 3 h before RI, the source heights of the NBE discharges increased continuously from ∼10 to ∼14 km (Fig. 3). This observation suggests that strengthening updrafts in the eyewall lofted the charge regions to progressively higher altitudes, similar to observations of intensifying severe continental storms (Reap and MacGorman 1989) and tornadic storms (Hamlin and Harlin 2002). After the continuous increase in NBE discharge heights, an abrupt decrease in heights and a sharp increase in CG flash rate occurred during the onset of the period of maximum intensity. This behavior is, again, similar to that of continental storms in which a descent of precipitation cores, and therefore charge regions, is often followed by a sharp increase in CG flash rate (e.g., Carey and Rutledge 1998).

Fig. 3.

Time–height plot of NBEs for Hurricane Rita with minimum central pressure overlaid as in Fig. 2a. The black ovals labeled A through D highlight lightning bursts analyzed in more detail in Fig. 4. The solid black line represents the approximate best-fit slope of the NBE discharge heights as a function of time.

Fig. 3.

Time–height plot of NBEs for Hurricane Rita with minimum central pressure overlaid as in Fig. 2a. The black ovals labeled A through D highlight lightning bursts analyzed in more detail in Fig. 4. The solid black line represents the approximate best-fit slope of the NBE discharge heights as a function of time.

Another new result of this study is the mapping–tracking of individual eyewall CEs by lightning observations. The time evolution of the 3D lightning locations was examined in Figs. 4a–d, in a sequence of short time intervals for the four NBE burst episodes labeled A–D in Fig. 3. All four CEs rotated in a counterclockwise fashion along the eyewall. The lifetime of the four episodes ranged from approximately 12 to 40 min. Horizontal projections of CG flashes made for the same short time intervals did not reveal such well-defined patterns (not shown). Although, admittedly, the extent of the heights for each case is comparable to the range of the systematic measurement uncertainty of the height calculation, the trend of relative heights from NBE to NBE does show an increase during the earlier half of the life cycles followed by a decrease, at least for the cases shown in Figs. 4c,d. This NBE lightning behavior is similar to what has been reported in continental storms for total IC flashes (e.g., MacGorman et al. 1989; Carey and Rutledge 1998): lofting of the charges by strengthening of the updraft and hence discharge heights followed by the weakening of the updraft and concomitant descent of the reflectivity core carrying the charges. If this NBE lightning behavior is indeed real then the individual CEs were almost certainly responsible for contributing to the general increase in the NBE discharge heights shown in Fig. 3. This suggests that NBEs can be used to map and track the evolution of CEs within intensifying hurricanes, and provide important information to modelers and operational forecasters on the small-scale convective structure/state of the eyewall, and hence potential imminent intensity changes of the storm.

Fig. 4.

Horizontal projections of NBEs, labeled A to D as in Fig. 3, during selected time intervals specified next to each of the four labels. The flashes were color coded in time for each event, where blue represents early in time and red later in time. The circle in the black line shows the estimated location and size of the eye of Rita based on NHC data. In panel B, the lightning locations disagree with the NHC’s storm center location, and the dashed circle indicates the location inferred from the lightning data. Although the exact reason(s) for this location disagreement is (are) unknown, it is likely that the latter might have been caused by either or both LASA or NHC position errors. (right) Corresponding temporal evolution of the heights of the IC discharges.

Fig. 4.

Horizontal projections of NBEs, labeled A to D as in Fig. 3, during selected time intervals specified next to each of the four labels. The flashes were color coded in time for each event, where blue represents early in time and red later in time. The circle in the black line shows the estimated location and size of the eye of Rita based on NHC data. In panel B, the lightning locations disagree with the NHC’s storm center location, and the dashed circle indicates the location inferred from the lightning data. Although the exact reason(s) for this location disagreement is (are) unknown, it is likely that the latter might have been caused by either or both LASA or NHC position errors. (right) Corresponding temporal evolution of the heights of the IC discharges.

To further illustrate the relationship between NBE bursts and intense convection, in situ airborne horizontal Doppler radar scans (Fig. 5) were compared with the corresponding NBE locations in Fig. 4c. Because no aircraft reconnaissance missions were flown at the time of Figs. 4a,b,d, similar relationships could not be established. The 30-dBZ echo threshold was chosen because past observational studies on tropical convection (e.g., Petersen et al. 1999) have shown that 30 dBZ was well correlated with lightning activity and significant hydrometeor mass. Both Figs. 4 and 5 show a good agreement between the time-evolving NBE locations and the rotating convective cell around the eyewall. Because low-level reflectivity maxima are not always spatially collocated with updraft maximum aloft, especially in the vertically sheared environment of the hurricane eyewall, three-dimensional renderings of radar reflectivity (Figs. 6a,b) were constructed during the time range shown in Fig. 5. Clearly, the northern portion of the eyewall, where the NBE flashes and maximum radar reflectivity at low levels were observed in Figs. 4a and 5, respectively, exhibits deeper radar echoes than the southern eyewall. Hence, Figs. 4a6 support the idea that NBEs are associated with deep convective updrafts in hurricane eyewalls.

Fig. 5.

(top) Rita’s time and spatial evolution of NBEs compared with (bottom) in situ lower fuselage Doppler radar reflectivity from aircraft reconnaissance. The four arrows labeled 1–4 corresponds to the four arrows shown in each of the radar reflectivity images. (Data courtesy of the NOAA/Hurricane Research Division.)

Fig. 5.

(top) Rita’s time and spatial evolution of NBEs compared with (bottom) in situ lower fuselage Doppler radar reflectivity from aircraft reconnaissance. The four arrows labeled 1–4 corresponds to the four arrows shown in each of the radar reflectivity images. (Data courtesy of the NOAA/Hurricane Research Division.)

Fig. 6.

Three-dimensional renderings of the 30-dBZ isosurface from in situ aircraft tail Doppler radar data at 1915 UTC 21 Sep for Hurricane Rita: (a) side view and (b) top view. The height above sea level was overlaid onto the isosurface with legends shown on the right of (a). Cardinal axes were added in (b) for clarity. (Data courtesy of the NOAA/HRD.)

Fig. 6.

Three-dimensional renderings of the 30-dBZ isosurface from in situ aircraft tail Doppler radar data at 1915 UTC 21 Sep for Hurricane Rita: (a) side view and (b) top view. The height above sea level was overlaid onto the isosurface with legends shown on the right of (a). Cardinal axes were added in (b) for clarity. (Data courtesy of the NOAA/HRD.)

Figure 6b also indicates that Rita’s eyewall exhibits a small vertical 30-dBZ reflectivity slope (or tilt, defined from the vertical axis). To further illustrate the lack of slope in Rita’s eyewall a three-dimensional projection of total lightning (CG and NBE) was constructed during a 1.5-h time period of intense lightning activity during RI (Fig. 7, which is near the time of Fig. 6) and shows that there is almost no radial displacement of the CG and NBE flash locations. This also suggests that the gross charge structure of Rita (or more precisely, of the CEs embedded in its eyewall) was almost upright.

Fig. 7.

Three-dimensional projection of NBE flashes (red) and CG flashes (blue) between 1615 and 1745 UTC 21 Sep 2005, during a period of intense lightning activity in the eyewall of Hurricane Rita. Vertical projection lines for IC flashes are shown to highlight the vertical displacement of NBEs with respect to the CG flashes.

Fig. 7.

Three-dimensional projection of NBE flashes (red) and CG flashes (blue) between 1615 and 1745 UTC 21 Sep 2005, during a period of intense lightning activity in the eyewall of Hurricane Rita. Vertical projection lines for IC flashes are shown to highlight the vertical displacement of NBEs with respect to the CG flashes.

4. Hurricane Katrina results

Similar to Rita, Katrina produced IC and CG lightning bursts during the onset of its first two periods of deepening: at 1500 UTC 26 August and at 1200 UTC 27 August (Fig. 8). An interesting result can be drawn from the timing of the first IC burst of Katrina when overlaid upon the pressure traces from the NHC best-track and 3-hourly public advisories datasets. When compared to the pressure trace estimate from the public advisories, the IC lightning exhibits two distinct outbursts on 26 August, which are coincident with the first two periods of deepening (Fig. 8). The CG lightning time series from its part shows an outburst during the first deepening period and before the onset of the second deepening stage, which was also seen before the RI of Rita. When considering the standard best-track pressure trace we see that the first two deepening periods of Katrina are not captured. Based on this result and those obtained for Rita, it seems that the pressure trace estimate from the NHC public advisories is more consistent with the timing of the IC outbursts.

Fig. 8.

As in Fig. 2a, but for Hurricane Katrina.

Fig. 8.

As in Fig. 2a, but for Hurricane Katrina.

Despite the longer time period (i.e., a 3-day period for Katrina compared to about 1 day for Rita) and its larger size, the eyewall of Katrina produced less lightning in total than Rita: LASA detected 1430 eyewall lightning flashes (ICs + CGs) for Rita, compared to 684 for Katrina. A recent study from Squires and Businger (2008), based solely on CG detection, hypothesized that the larger, and therefore more tilted, eyewall (Stern and Nolan 2009) of Katrina caused the main charge regions to be radially displaced from each other in the vertical, in turn reducing the area of strong electric field and hence lightning activity in the eyewall. This hypothesis is in line with Black and Hallett’s (1999) seminal work showing that in hurricane eyewalls the ratio of ice to supercooled water, a key ingredient for noninductive collisional charge separation with riming graupel (e.g., Takahashi 1978; Saunders and Peck 1998), increases radially outward.

The time–height plot of NBE discharge heights for Katrina in Fig. 9 shows evidence of an increase in height of NBE events during its periods of deepening. During those outbursts, isolated convective burst events can also be seen (again, highlighted by ovals in Fig. 9). Also, similar to Rita, Katrina did exhibit a sharp increase in CG flash rate during its period of maximum intensity (Fig. 8). A similar CG and NBE behavior was seen for Charley (not shown). Although this behavior is reported for only three hurricane cases here, these findings suggest that the occurrence and increase in height of NBE bursts in the eyewall might occur in other intensifying hurricanes.

Fig. 9.

As in Fig. 3, but for Hurricane Katrina.

Fig. 9.

As in Fig. 3, but for Hurricane Katrina.

Horizontal projections of IC lightning flashes during short time intervals similar to those shown in Fig. 4 were also constructed in Fig. 10 for Katrina for the two NBE outbursts occurring between about 0900 and 1200 UTC 27 August 2005 (Fig. 9). Similar plots for Katrina at earlier times were not included because of unfavorable LASA network geometry, which, as mentioned in the previous section, would yield unreliable location solutions. Figure 10 shows that the lightning activity followed the storm motion (which was nearly due west at that time) and persisted for about a 3-h period south of the eyewall. This presence of persistent strong convection to the south of Katrina’s eyewall was confirmed by radar and satellite imagery (not shown) and could have likely been attributed to external forcing such as (southerly) wind shear (e.g., Corbosiero and Molinari 2002; Rogers et al. 2003).

Fig. 10.

As in Fig. 4, but for Hurricane Katrina for the two bursts highlighted in Fig. 9 between about 0900 and 1300 UTC 27 Aug 2005.

Fig. 10.

As in Fig. 4, but for Hurricane Katrina for the two bursts highlighted in Fig. 9 between about 0900 and 1300 UTC 27 Aug 2005.

It is interesting to note that the lightning in Fig. 10 did not show evidence of rotation around the eyewall like it did for Rita. As Katrina moved westward the NBE discharge heights did show an increase of about 2–3 km indicating that the convection in the southern eyewall became deeper with time (Fig. 9). The above findings show that the behavior of the convection mapped by the IC lightning between Rita and Katrina was different, despite the storms having relatively similar intensities (∼950 hPa) during the time periods considered.

5. Discussion and conclusions

This work illustrates various new aspects of lightning activity observed within Katrina and Rita—two rapidly intensifying category 5 hurricanes. Based on CG flash data, previous studies have highlighted that rapidly intensifying hurricanes could produce abundant eyewall lightning (e.g., Molinari et al. 1994). The current work supports these previous findings and also demonstrates that during the onset of RI, CG, and IC flashes together can be used to provide important insight into the structure of the storm, such as eye evolution and eyewall convective state.

For Katrina and Rita, the discharge heights of NBEs exhibited an increase in height during periods of deepening, which was then followed by a rapid decrease in height after the RI period. The former suggests a continuous lofting of charged particles by eyewall updrafts. In turn, the decrease in NBE height indicates a collapse of charged particles, and hence, updrafts. Furthermore, the NBEs in the eyewall of Rita were found to follow the rotating convective cells around the eyewall, and could be used in this case to track the temporal and spatial evolution of individual convective cells. This shows that for Rita, NBE lightning provides a detailed insight into the dynamics of the eyewall that are otherwise difficult to capture.

Given that NBE observations could be used to resolve the evolution of strong CEs, a proxy for latent heat (i.e., NBE discharges) can be formulated to initialize hurricane models. Additionally, since lightning discharges could be observed continuously over long time periods, the data could be readily incorporated into four-dimensional data assimilation procedures to help determine the convective state of the hurricane eyewall. The assimilation of the small-scale CEs also raises important questions with regard to the community’s ability to predict hurricane intensification. Chief among them is whether hurricane models must accurately resolve the evolution of the CEs or just capture the integrated impact of these events. If the former is indeed required to accurately model hurricane intensification, then an even more vexing question must be addressed: What process(es) is (are) responsible for forcing the CEs? A recent review article by Houze (2010) describes the possible mechanisms put forth to date, which are believed to account for the formation of eyewall CEs. Since those CEs are embedded into a vorticity-rich environment provided by the primary vortex, those updrafts were often shown to exhibit strong rotation as measured by helicity (e.g., Molinari and Vollaro 2008) and were hence coined the term vortical hot towers (VHTs). Eyewall VHTs were in some cases associated with the presence of persistent weak-to-moderate ambient vertical wind shear (Rogers et al. 2003; Molinari and Vollaro 2008, 2010). In the case of Rita, however, very little-to-no shear was present during its RI period (Knabb et al. 2005) and therefore other external and/or internal physical mechanisms must have been at play to account for the formation of the CEs mapped by the LASA NBE flashes.

Future work should also investigate whether the lightning behavior documented herein for NBEs could be applied to weaker tropical systems, especially tropical storms rapidly intensifying into hurricanes (e.g., Hurricane Humberto in 2007). This last question is particularly intriguing as the recent study from Molinari and Vollaro (2010), which described the RI of Tropical Storm Gabrielle (2001) into a hurricane during a few hours prior to making landfall revealed a strong convective cell developing downshear left of the center and spiraling cyclonically inward during the period of RI as seen in Rita. They found that this CE had abundant CG lightning activity associated with it (as detected by the NLDN). The Tropical Rainfall Measuring Mission (TRMM) precipitation radar also indicated an area of 35 dBZ at the 12-km elevation threshold, which would likely be associated with significant IC (NBE) flash activity (e.g., Petersen et al. 1999; Black and Hallett 1999).

Since the lightning shows that CEs were associated in each case with RI of the hurricane, it may be key to advance our understanding on how these deep thunderstorms form and to determine what internal or external forcings are at play. Because of the generally transient nature of these events (∼30 min to 3 h), more in situ and remote sensing data taken with both higher detection efficiencies and higher temporal resolution will prove itself useful to help unveil the physical mechanisms behind the formation of CEs and hence, behind the rapid intensification of some hurricanes.

Acknowledgments

This work was supported by the U.S. Department of Energy and the LDRD program of the Los Alamos National Laboratory. The authors would also like to thank Dr. Edward “Ted” Mansell, Dr. Don MacGorman, Dr. Conrad Ziegler, and Dr. Chris Jeffery for providing helpful suggestions on an earlier version of the manuscript. The authors would also like to thank Dr. Dave Vollaro and one anonymous reviewer for their helpful comments. Thanks also go out to Steve Guimond for providing the 3D radar data used to create Fig. 6.

REFERENCES

REFERENCES
Black
,
R. A.
, and
J.
Hallett
,
1999
:
Electrification of the hurricane
.
J. Atmos. Sci.
,
56
,
2004
2028
.
Carey
,
L. D.
, and
S. A.
Rutledge
,
1998
:
Electrical and multiparameter radar observations of a severe hailstorm
.
J. Geophys. Res.
,
103
(
D12
),
13 979
14 000
.
Corbosiero
,
K. L.
, and
J.
Molinari
,
2002
:
The effects of vertical wind shear on the distribution of convection in tropical cyclones
.
Mon. Wea. Rev.
,
130
,
2110
2123
.
DeMaria
,
M.
, and
R. T.
DeMaria
,
2009
:
Applications of lightning observations to tropical cyclone intensity forecasting
.
Preprints, 16th Conf. on Satellite Meteorology and Oceanography, Phoenix, AZ, Amer. Meteor. Soc., 1.3. [Available online at http://ams.confex.com/ams/89annual/techprogram/paper_145745.htm.]
Eack
,
K. B.
,
2004
:
Electrical characteristics of narrow bipolar events
.
Geophys. Res. Lett.
,
31
,
L20102
,
doi:10.1029/2004GL021117
.
Fierro
,
A. O.
, and
J. M.
Reisner
,
2011
:
High-resolution simulation of the electrification and lightning of Hurricane Rita during the period of rapid intensification
.
J. Atmos. Sci.
,
68
,
477
494
.
Fierro
,
A. O.
,
L. M.
Leslie
,
E. R.
Mansell
,
J. M.
Straka
,
D. R.
MacGorman
, and
C.
Ziegler
,
2007
:
A high resolution simulation of the microphysics and electrification in an idealized hurricane-like vortex
.
Meteor. Atmos. Phys.
,
98
,
13
33
,
doi:10.1007/s00703-006-0237-0
.
Guimond
,
S. R.
,
G. M.
Heymsfield
, and
F. J.
Turk
,
2010
:
Multiscale observations of Hurricane Dennis (2005): The effects of hot towers on rapid intensification
.
J. Atmos. Sci.
,
67
,
633
654
.
Hamlin
,
T.
, and
J.
Harlin
,
2002
:
Tornado signatures and precursor activity from 3-D lightning mapping observations
.
Preprints, 21st Conf. on Severe Local Storms, San Antonio, TX, Amer. Meteor. Soc., 11A.5. [Available online at http://ams.confex.com/ams/SLS_WAF_NWP/techprogram/paper_47438.htm.]
Hamlin
,
T.
,
T. E.
Light
,
X. M.
Shao
,
K. B.
Eack
, and
J. D.
Harlin
,
2007
:
Estimating lightning channel characteristics of positive narrow bipolar events using intra-channel current reflection signatures
.
J. Geophys. Res.
,
112
,
D14108
,
doi:10.1029/2007JD008471
.
Hendricks
,
E. A.
,
M. T.
Montgomery
, and
C. A.
Davis
,
2004
:
The role of “vortical” hot towers in the formation of Tropical Cyclone Diana (1984)
.
J. Atmos. Sci.
,
61
,
1209
1232
.
Houze
,
R. A.
,
2010
:
Clouds in tropical cyclones
.
Mon. Wea. Rev.
,
138
,
293
344
.
Jacobson
,
A. R.
, and
M. J.
Heavner
,
2005
:
Comparison of narrow bipolar events with ordinary lightning as proxies for severe convection
.
Mon. Wea. Rev.
,
133
,
1144
1154
.
Jacobson
,
A. R.
,
R.
Holzworth
,
J.
Harlin
,
R.
Dowden
, and
E.
Lay
,
2006
:
Performance assessment of the World Wide Lightning Location Network (WWLLN) using the Los Alamos Sferic Array (LASA) as ground truth
.
J. Atmos. Oceanic Technol.
,
23
,
1082
1092
.
Jorgensen
,
D. F.
,
1984
:
Mesoscale and convective-scale characteristics of mature hurricanes. Part I: General observations by research aircraft
.
J. Atmos. Sci.
,
41
,
1268
1286
.
Kaplan
,
J.
, and
M.
DeMaria
,
2003
:
Large-scale characteristics of rapidly intensifying tropical cyclones in the North Atlantic basin
.
Wea. Forecasting
,
18
,
1093
1108
.
Kaplan
,
J.
,
M.
DeMaria
, and
J. A.
Knaff
,
2010
:
A revised tropical cyclone rapid intensification index for the Atlantic and eastern North Pacific basins
.
Wea. Forecasting
,
25
,
220
241
.
Kelley
,
O. A.
,
J.
Stout
, and
J. B.
Halverson
,
2004
:
Tall precipitation cells in tropical cyclone eyewalls are associated with tropical cyclone intensification
.
Geophys. Res. Lett.
,
31
,
L24112
,
doi:10.1029/2004GL021616
.
Knabb
,
R. D.
,
D. P.
Brown
, and
J. R.
Rhome
,
2005
:
Tropical cyclone report: Hurricane Rita, 18–26 September 2005
.
Tech. Rep., NOAA/National Hurricane Center, 33 pp. [Available online at http://www.nhc.noaa.gov/2005atlan.shtml.]
Landsea
,
C. W.
,
1993
:
A climatology of intense (or major) Atlantic hurricanes
.
Mon. Wea. Rev.
,
121
,
1703
1713
.
Landsea
,
C. W.
, and
Coauthors
,
2004
:
The Atlantic hurricane database re-analysis project: Documentation for the 1851-1910 alterations and additions to the HURDAT database
.
Hurricanes and Typhoons: Past, Present and Future, R. J. Murname and K.-B. Liu, Eds., Columbia University Press, 177–221
.
Lay
,
E. H.
,
A. R.
Jacobson
,
R. H.
Holzworth
,
C. J.
Rodger
, and
R. L.
Dowden
,
2007
:
Local time variation in land/ocean lightning count rates as measured by the World Wide Lightning Location Network
.
J. Geophys. Res.
,
112
,
D13111
,
doi:10.1029/2006JD007944
.
Le Vine
,
D. M.
,
1980
:
Source of the strongest RF radiation from lightning
.
J. Geophys. Res.
,
85
,
4091
4095
.
MacGorman
,
D. R.
,
D. W.
Burgess
,
V.
Mazur
,
W. D.
Rust
,
W. L.
Taylor
, and
B. C.
Johnson
,
1989
:
Lightning rates relative to tornadic storm evolution on 22 May 1981
.
J. Atmos. Sci.
,
46
,
221
251
.
Marks
,
F.
, and
L. K.
Shay
,
1998
:
Landfalling tropical cyclones: Forecast problems and associated research opportunities
.
Bull. Amer. Meteor. Soc.
,
79
,
305
323
.
Molinari
,
J.
, and
D.
Vollaro
,
2008
:
Extreme helicity and intense convective towers in Hurricane Bonnie
.
Mon. Wea. Rev.
,
136
,
4355
4372
.
Molinari
,
J.
, and
D.
Vollaro
,
2010
:
Rapid intensification of sheared tropical storm
.
Mon. Wea. Rev.
,
138
,
3869
3885
.
Molinari
,
J.
,
P. K.
Moore
,
V. P.
Idone
,
R. W.
Henderson
, and
A. B.
Saljoughy
,
1994
:
Cloud-to-ground lightning in Hurricane Andrew
.
J. Geophys. Res.
,
99
,
16 665
16 676
.
Molinari
,
J.
,
P.
Moore
, and
V.
Idone
,
1999
:
Convective structure of hurricanes as revealed by lightning locations
.
Mon. Wea. Rev.
,
127
,
520
534
.
Montgomery
,
M. T.
,
M. E.
Nicholls
,
T. A.
Cram
, and
A. B.
Saunders
,
2006
:
A vortical hot tower route to tropical cyclogenesis
.
J. Atmos. Sci.
,
63
,
355
386
.
Neumann
,
C. J.
,
1994
:
An update to the National Hurricane Center “Track Book.”
.
Minutes of the 48th Interdepartmental Conference, NOAA/Office of Federal Coordinator for Meteorological Services and Supporting Research, Miami, FL, A-47–A-53
.
Petersen
,
W. A.
,
R. C.
Cifelli
,
S. A.
Rutledge
,
B. S.
Ferrier
, and
B. F.
Smull
,
1999
:
Shipborne dual-Doppler operations and observations during TOGA COARE
.
Bull. Amer. Meteor. Soc.
,
80
,
81
97
.
Powell
,
M. D.
,
1990
:
Boundary layer structure and dynamics in outer hurricane rainbands. Part II: downdraft modification and mixed layer recovery
.
Mon. Wea. Rev.
,
118
,
918
938
.
Price
,
C.
,
M.
Asfur
, and
Y.
Yair
,
2009
:
Maximum hurricane intensity preceded by increase in lightning frequency
.
Nat. Geosci.
,
2
,
329
332
,
doi:10.1038/NGEO477
.
Reap
,
R. M.
, and
D. R.
MacGorman
,
1989
:
Cloud-to-ground lightning: Climatological characteristics and relationships to model fields, radar observations, and severe local storms
.
Mon. Wea. Rev.
,
117
,
518
535
.
Reasor
,
P. D.
,
M. D.
Eastin
, and
J. F.
Gamache
,
2009
:
Rapidly intensifying Hurricane Guillermo (1997). Part I: Low-wavenumber structure and evolution
.
Mon. Wea. Rev.
,
137
,
603
631
.
Rison
,
W.
,
R. J.
Thomas
,
P. R.
Krehbiel
,
T.
Hamlin
, and
J.
Harlin
,
1999
:
A GPS-based three-dimensional lightning mapping system: Initial observations in central New Mexico
.
Geophys. Res. Lett.
,
26
,
3573
3576
.
Rodgers
,
E. B.
,
W. S.
Olson
,
V. M.
Karyampudi
, and
H. F.
Pierce
,
1998
:
Satellite-derived latent heating distribution and environmental influences in Hurricane Opal (1995)
.
Mon. Wea. Rev.
,
126
,
1229
1247
.
Rogers
,
R.
,
S.
Chen
,
J.
Tenerelli
, and
H.
Willoughby
,
2003
:
A numerical study of the impact of vertical shear on the distribution of rainfall in Hurricane Bonnie (1998)
.
Mon. Wea. Rev.
,
131
,
1577
1599
.
Saunders
,
C. P. R.
, and
S. L.
Peck
,
1998
:
Laboratory studies of the influence of the rime accretion rate on charge transfer during crystal/graupel collisions
.
J. Geophys. Res.
,
103
(
D12
),
13 949
13 956
.
Shao
,
X. M.
, and
A. R.
Jacobson
,
2009
:
Model simulation of very low-frequency and low-frequency lightning signal propagation over intermediate ranges
.
IEEE Trans. EMC
,
51
,
519
525
.
Shao
,
X. M.
, and
Coauthors
,
2005
:
Katrina and Rita were lit up with lightning
.
Eos, Trans. Amer. Geophys. Union
,
86
,
doi:10.1029/2005EO420004
.
Shao
,
X. M.
,
M.
Stanley
,
A.
Regan
,
J.
Harlin
,
M.
Pongratz
, and
M.
Stock
,
2006
:
Total lightning observations with the new and improved Los Alamos Sferic Array (LASA)
.
J. Atmos. Oceanic Technol.
,
23
,
1273
1288
.
Simpson
,
J.
,
J. B.
Halverson
,
B. S.
Ferrier
,
W. A.
Petersen
,
R. H.
Simpson
,
R.
Blakeslee
, and
S. L.
Durden
,
1998
:
On the role of “hot towers” in tropical cyclone formation
.
Meteor. Atmos. Phys.
,
67
,
15
35
.
Smith
,
D. A.
, and
Coauthors
,
1999
:
A distinct class of isolated intracloud lightning discharges and their associated radio emissions
.
J. Geophys. Res.
,
104
,
4189
4212
.
Smith
,
D. A.
,
M. J.
Heavner
,
A. R.
Jacobson
,
X. M.
Shao
,
R. S.
Massey
,
R. J.
Sheldon
, and
K. C.
Wiens
,
2004
:
A method for determining intracloud lightning and ionospheric heights from VLF/LF electric field records
.
Radio Sci.
,
39
,
RS1010
,
doi:10.1029/2002RS002790
.
Solorzano
,
N. N.
,
J. N.
Thomas
, and
R. H.
Holzworth
,
2008
:
Global studies of tropical cyclones using the World Wide Lightning Location Network
.
Preprints, Third Conf. on the Meteorological Applications of Lightning Data, New Orleans, LA, Amer. Meteor. Soc., 1.4. [Available online at http://wwlln.net/publications/Solorzano_AMS_29FEB2008.pdf.]
Squires
,
K.
, and
S.
Businger
,
2008
:
The morphology of eyewall lightning outbreaks in two category 5 hurricanes
.
Mon. Wea. Rev.
,
136
,
1706
1726
.
Stern
,
D. P.
, and
D. S.
Nolan
,
2009
:
Reexamining the vertical structure of tangential winds in tropical cyclones: Observations and theory
.
J. Atmos. Sci.
,
66
,
3579
3600
.
Takahashi
,
T.
,
1978
:
Riming electrification as a charge generation mechanism in thunderstorms
.
J. Atmos. Sci.
,
35
,
1536
1548
.
Thomas
,
J. N.
,
N.
Solorzano
,
S. A.
Cummer
, and
R. H.
Holzworth
,
2010
:
Polarity and energetics of inner core lightning in three intense North Atlantic hurricanes
.
J. Geophys. Res.
,
115
,
A00E15
,
doi:10.1029/2009JA014777
.
Uman
,
M. A.
,
1987
:
The Lightning Discharge
.
Academic Press, 377 pp
.
Wang
,
Y.
,
2009
:
How do outer spiral rainbands affect tropical cyclone structure and intensity?
J. Atmos. Sci.
,
66
,
1250
1273
.
Wiens
,
K. C.
,
S. A.
Rutledge
, and
S. A.
Tessendorf
,
2005
:
The 29 June 2000 supercell observed during STEPS. Part II: Lightning and charge structure
.
J. Atmos. Sci.
,
62
,
4151
4177
.
Willett
,
J. C.
,
J. C.
Bailey
, and
E. P.
Krider
,
1989
:
A class of unusual lightning electric field waveforms with very strong high-frequency radiation
.
J. Geophys. Res.
,
94
,
16 255
16 267
.
Williams
,
E. R.
,
1989
:
The tripole structure of thunderstorms
.
J. Geophys. Res.
,
94
,
13 151
13 167
.
Williams
,
E. R.
,
1995
:
Meteorological aspects of thunderstorms
.
Handbook of Atmospheric Electrodynamics, H. Volland, Ed., Vol. 1, CRC Press, 27–60
.