AE33A-0174 1340h
Automatic detection of lightning generated whistlers using data from the QuakeSat I satellite.
Using a signal detection technique based on the dispersion characteristics of a typical lightning generated whistler, we analyze the data recorded on the magnetometer of the QuakeSat I satellite, at an altitude of 840 km. Results show that our analysis technique is able to detect whistler signals that are well below the noise threshold and cannot be visually detected in frequency-time spectrograms. As an example, we pick several lightning-generated whistler events from our data and quantify the expected delay and damping of the signals using a ray tracing model and data from the National Lightning Detection Network, to show good agreement with observations.
AE33A-0175 1340h
Comparison of World Wide Lightning Location Network and FORTE lightning data
The World Wide Lightning Location (WWLL) system [Dowden et al., 2002] provides real time lightning locations globally by measuring the very low frequency (VLF) radiation emanating from lightning discharges. The location accuracy and efficiency of the WWLL have been estimated for certain regions of the globe [{\it Lay et al.}, 2004; {\it Rodger et al.}, 2004]. The FORTE satellite has monitored lightning within its field-of-view using radiofrequency and optical sensors [{\it Jacobson et al.}, 2000; {\it Kirkland et al.}, 2001; {\it Suszcynsky et al.}, 2000]. The signals recorded by FORTE have some capability to identify the lightning discharge process, e.g. ground stroke, intracloud stroke, or Narrow Bipolar Event [{\it Jacobson and Light}, 2003; {\it Jacobson and Shao}, 2002]. We will report on initial comparisons between WWLL and FORTE observations, conducted to gain more insights on the spatial and temporal accuracy of the WWLL as well as its ability to detect lightning discharges of various types. Dowden, R.L., J.B. Brundell, and C.J. Rodger, {\it J. Atmos. Sol. Terr. Phys., 64}, 817-830, 2002. Jacobson, A.R., K.L. Cummins, M. Carter, P. Klingner, D. Roussel-Dupre, and S.O. Knox, {\it J. Geophys. Res., 105} (D12), 15,653, 2000. Jacobson, A.R., and T.E.L. Light, {\it J. Geophys. Res., 108} (D9), 4266, doi:10.1029/2002JD002613, 2003. Jacobson, A.R., and X.-M. Shao, {\it J. Geophys. Res., 107} (D22), 4661, doi:10.1029/2001JD001542, 2002. Kirkland, M.W., D.M. Suszcynsky, J.L.L. Guillen, and J.L. Green, {\it J. Geophys. Res., 106} (D24), 33,499-33,509, 2001. Lay E.H., R.H. Holzworth, C.J. Rodger, J.N. Thomas, O. Pinto Jr., R.L. Dowden, {\it Geophys. Res. Lett., 31}, doi:10.1029/2003GL018882, 2004. Rodger, C. J., J.B. Brundell, R.L. Dowden, and N.R. Thomson, {\it Annales Geophys. 22}, 747-758, 2004. Suszcynsky, D.M., M.W. Kirkland, A.R. Jacobson, R.C. Franz, S.O. Knox, J.L.L. Guillen, and J.L. Green, {\it J. Geophys. Res., 105} (D2), 2191-2201, 2000.
AE33A-0176 1340h
A Preliminary ZEUS Lightning Location Error Analysis Using a Modified Retrieval Theory
The ZEUS long-range VLF arrival time difference lightning detection network now covers both Europe and Africa, and there are plans for further expansion into the western hemisphere. In order to fully optimize and assess ZEUS lightning location retrieval errors and to determine the best placement of future receivers expected to be added to the network, a software package is being developed jointly between the NASA Marshall Space Flight Center (MSFC) and the University of Nevada Las Vegas (UNLV). The software package, called the ZEUS Error Analysis for Lightning (ZEAL), will be used to obtain global scale lightning location retrieval error maps using both a monte carlo approach and chi-squared curvature matrix theory. At the core of ZEAL will be an implementation of an Iterative Oblate (IO) lightning location retrieval method recently developed at MSFC [Koshak and Solakiewicz, TOA Lightning Location Retrieval on Spherical and Oblate Spheroidal Earth Geometries, J. Atmos. and Oceanic. Tech., 187-199, 2001.] The IO method will be appropriately modified to account for variable wave propagation speed, and the new retrieval results will be compared with the current ZEUS retrieval algorithm to assess potential improvements. In this preliminary ZEAL work effort, we defined 5000 source locations evenly distributed across the Earth. We then used the existing (as well as potential future ZEUS sites) to simulate arrival time data between source and ZEUS site. A total of 100 sources were considered at each of the 5000 locations, and timing errors were selected from a normal distribution having a mean of 0 seconds and a standard deviation of 20 microseconds. This simulated "noisy" dataset was analyzed using the IO algorithm to estimate source locations. The exact locations were compared with the retrieved locations, and the results are summarized via several color-coded "error maps."
AE33A-0177 1340h
The 2002-2003 Upgrade of the U.S. National Lightning Detection Network
Beginning in the spring of 2002, the U.S. National Lightning Detection Network (NLDNT) has been undergoing a system-wide upgrade. The objectives of this upgrade are to increase sensor reliability and reduce maintenance costs, to provide enhanced detection efficiency and location accuracy on the boundaries of the network, and to detect a fraction of the of cloud discharges. The original (1995) time-of-arrival (TOA) LPATS sensors and early IMPACT (DF + TOA) sensors have been replaced by new IMPACT ESP sensors, however two LPATS sensors have yet to be replaced. An additional 7 sensor sites have been installed and one more is planned. The new IMPACT ESP sensors provide both accurate TOA and direction information and increased sensitivity. The estimated stroke detection efficiency (DE) is now in the range of 60-80%, and the overall modeled cloud-to-ground flash DE exceeds 90% throughout the U.S. This DE estimate is no longer limited to events above 5 kA. This paper will provide the research community with up-to-date information about the on-going improvements to the NLDN, a comparison of the modeled and measured detection efficiency both before and after the upgrade, and a discussion of the small positive reports. The effect of the upgrade on CG peak current distributions and flash multiplicities will also be presented.
AE33A-0178 1340h
Performance Validation of the 2002-2003 Upgrade of the U.S. National Lightning Detection Network
Beginning in the spring of 2002, the U.S. National Lightning Detection Network (NLDNT) has been undergoing a system-wide upgrade. In conjunction with this upgrade, independent field evaluations have been carried out in Southern Arizona and in Oklahoma/Texas in 2003 and 2004, and at the International Center for Lightning Research and Testing (ICLRT) at Camp Blanding, Florida, in 2001-2003. Data from these studies have been used to quantify the performance of the NLDN and to evaluate the long-standing problem of the NLDN misclassifying cloud discharges as small positive cloud-to-ground strokes. The overall flash DE in Southern Arizona after the upgrade is 95%, based on 2290 video-measured return strokes, and the flash DE in Oklahoma and Texas is now about 93%, based on 734 video-measured strokes. Video observations of over 500 NLDN events that were classified as small positive strokes indicate that nearly all positive reports with an estimated peak current below 10 kA are cloud discharges, and most positive reports with estimated peak currents above 20 kA are CG strokes. The positive events between 10-20 kA (about 10% of the positive discharges and about 1% of all discharges) are a mix of cloud discharges and CG strokes. There was no indication of significant misclassification of small negative discharges. The ICLRT studies suggest that the flash DE (based solely on subsequent strokes) near Gainesville, Florida, is 84%, a value that is consistent with NLDN flash DE in that area of 90% for natural lightning. The observed stroke DE at the ICLRT in 2003 was 69%, and the median location error was 450 meters. The median absolute peak current estimation error was about 20%.
AE33A-0179 1340h
The New LANL Sferic Array: Operation and Calibration
The new Los Alamos National Laboratory (LANL) sferic array was setup in a compact configuration in northern Florida in April, 2004 with additional stations to be deployed soon around the Great Plains. The array was operated for several months with 1 ms length triggers with the trigger-point normally centered in the data record. In August, the sensitivity was increased while the trigger length was shortened to 400 microseconds with a 80/320 microsecond pre/post-trigger. While even shorter trigger lengths could be used to drop the data rate (or increase the sensitivity for a constant data rate), this would compromise the detection of ionospheric reflections. The northern Florida stations were deployed on top of buildings of varying heights and widths. It was soon discovered that the sensors which were deployed on the tallest buildings were significantly more sensitive to sferics than those on the shorter ones, as would be expected due to the stronger focusing of electric field lines to taller structures. By comparing the amplitude of distant sferics between the stations, a relative gain calibration can be made. To obtain an absolute calibration, a flat-plate sensor with a known response will be located on flat ground near one of our sensors and sferic amplitudes will be compared between the sensors.
AE33A-0180 1340h
Locating Lightning With the New LANL Sferic Array
In the spring of 2004 Los Alamos National Laboratory deployed a new electric field change array (EDOTX) in north central Florida. Six of the eight stations are placed inside an area of about 150 km in diameter, producing high resolution locations of events within the network. The two distant stations allow for a wider field of coverage. The 1-500 kHz signal is digitized at 2 Mhz and the typical record length is 1 millisecond, with some events producing multiple back to back records. The previous array had record lengths of 8 milliseconds on average, and was deployed over a much larger area. The old method for 2-D geo-location was to cross-correlate the waveforms, get one time per record, and then use three or more stations to locate that event using a downhill simplex method. The new method takes a different approach. First the data is converted to power using a Hilbert transform, then individual peaks are selected and matched among the stations. Depending upon the complexity of the waveform and number of back to back records this method can produce multiple sets of times for a given event. Then for each set of times, if there are at least four stations, we obtain an initial guess for the location by a linear inversion, which is then passed to a least squares routine. The end result is more 2-D geo-located sources, each with an error measure, and the processing time is much faster compared to the old method. With this system it is possible to resolve separate cells within the storms, and even study the channel development of individual flashes.
AE33A-0181 1340h
New and Improved Los Alamos Sferic Array for Lightning Observations
In the spring of 2004, Los Alamos National Laboratory deployed eight fast field change sensors in northern-central Florida to observe the lightning activity in the surrounding areas. Six of the sensors formed a relatively dense array that covers an area of about 100km x 150km centered near Palatka. The other two sensors were located more remotely at Tampa and Tallahassee, over 200 km from the array center. Compared to our previous sferic array in Florida that consisted of five stations and was spread over the entire state, the current one was planned to provide more detailed lightning observations over the central array and at the same time to maintain the capability of long distance detection with the two remote stations. The electronics and data acquisition for the individual sensor were also improved substantially. The system is more stable and the noise level is significantly reduced. The trigger system is now software-based so that various and sophisticated trigger criteria can be easily implemented. The new system eliminated the "dead time" and is able to catch every pulse above the trigger threshold. Each station is equipped with an upgraded GPS receiver that time-tags the data at 50ns accuracy. The array has been operating continuously from April, 2004, and obtained a large amount of data for Hurricanes Charley and Frances. In this presentation, we will give an overview of the array and the observations. Detailed operation and results will be presented in other companion talks.
AE33A-0182 1340h
Preliminary Comparison of Narrow Bipolar Events With TRMM/TMI 85-GHz Ice-scattering Signature
The Edot array has provided ground-truth measurements, mainly in the Florida region, of lightning processes also observed by the FORTE satellite. Edot records and stores the entire wideband (1 Megasample/sec) sferic waveform. Waveform analysis allows identification of "Narrow Bipolar Events" or NBEs, a unique form of fast (~10 microsecond-duration) intracloud discharge that is not preceded by a leader phase. NBEs appear to be related to a class of extremely intense Very High Frequency (30-300 MHz) emissions that can be triggered upon using radio-receiver instruments in space. We will analyze the relationship of NBEs to the ice-scattering signature imagery provided by the TRMM/TMI microwave radiometer. The goal is to compare incidence and morphology between intense intracloud lightning and centers of deep, active convection.
AE33A-0183 1340h
Understanding the spatio-temporal characteristics of Thunderstorms in African by means of long range lightning detection
This study aims at studying the spatial and temporal variability and meteorology of Mesoscale Convective Systems (MCSs) in Africa. The research is facilitated by remotely sensed data that include instantaneous precipitation type and rain rate fields derived from TRMM Microwave Imager (TMI) observations, continuous cloud-to-ground lightning observations and coincident hourly satellite Infrared fields. The CG lightning data are based on observations from a newly deployed long-range detection network in Africa (named Zeus; sifnos.engr.uconn.edu). Zeus retrieval error in location and the detection efficiency will be evaluated using NASA LIS observations as reference. A tracking technique will then be applied to the IR and lightning fields to study the evolution of clouds and convective precipitation in the region. A convective-stratiform rain estimation technique will be developed (using PR as reference) and validated against in situ data.
http://sifnos.engr.uconn.edu
AE33A-0184 1340h
TRMM Observations of the Basic Relationship Between Ice Water Path and Lightning
Space-based lightning and radar observations are used to address the relationship between lightning and precipitation ice water mass (e.g., integrated amounts of graupel and hail), and the degree to which this relationship remains constant between global tropical ocean, coastal and continental convective regimes. Considering the close microphysical coupling between cloud electrification, lightning, and the ice phase, we hypothesize that regardless of regime there should be little difference in the relationship between convective ice-phase precipitation mass and lightning flash density (in contrast to the well documented variability between lightning and rainfall amount, where the microphysical coupling is often much weaker). To investigate our hypothesis we examined three years (1998-2000) of Tropical Rainfall Measurement Mission (TRMM) Lightning Imaging Sensor (LIS) flash density (FD) and TRMM Precipitation Radar (PR) reflectivity (Z) data for southern and northern hemisphere summer seasons (July-Aug., Dec.-Feb.). First, ice water contents (IWCs) were computed for PR range gates located at heights above the -10 C level using Z-IWC relationships based on exponential ice particle size distributions and ice particle densities adjusted as a function of precipitation type (convective or stratiform) and Z. Next, ice water paths (IWPs; ice mass in a column) were computed by vertically integrating IWC upward from the -10C level to radar echo top. Pixel-level IWP, FD and rainfall were subsequently averaged and gridded at a resolution of 0.5 x 0.5 degrees for comparison. When the grid-box IWPs were binned by FD and then globally averaged for land, ocean and coastal regimes, correlations between FD and IWP were found to be quite large (R$>$0.9) for all three regimes. Best-fit lines between FD and IWP for all three regimes were nearly identical, exhibiting differences in slope (a metric of "ice-yield" as opposed to "rain-yield") of less than 20% with nearly identical intercepts of ~0.01 kg/m^2 (an apparent lower threshold in IWP for achieving a detectable FD with LIS in a 0.5 x 0.5 Deg. grid cell). Appropriate power law fits between averaged FD and mean rainfall rate for all three regimes were also quite good (R$>$0.8); however, in contrast to the FD-IWP relationship. the rainfall-FD relationships by regime were markedly different. The results suggest that 1) when averaged over all oceans, land and coastal regimes there is no discernable difference in the relationship between ice-phase precipitation mass and FD between ocean, coastal and land regimes (in contrast to rainfall); and 2) to first order, the physical assumptions of precipitation-based charging and mixed phase precipitation development are robust, suggesting that lightning data may be a useful variable for inclusion in combined algorithms developed to retrieve precipitation ice water content (at least over the large scales examined herein).
AE33A-0185 1340h
Meteorological Forcing of Lightning over the Gulf Stream
From the perspective of both the National Lightning Detection Network and the Lightning Imaging Sensor onboard TRMM, it is well known that there is a local lightning maximum located over the Gulf Stream region off the eastern coast of the U.S. Flash densities are roughly 10 flashes/km2/month, comprable to flash rates over the adjacent continental land mass to the west. TRMM Precipitation Radar data also indicate characteristically "continental" convective vertical structure and ice water paths, consistent with the lightning signature. The diurnal cycle of lightning (and precipitation) over the Gulf Stream exhibits an extended nocturnal and early morning peak, reaching a maximum near 12 UTC (7 AM local). NLDN lightning data suggest that systems forming over land the previous day may undergo an extended period of resurgence in intensity and thus their ability to generate lightning as they pass over the warm waters of the Gulf Stream (28 C). Analysis of climatological variables in active periods of Gulf Stream lightning indicate the presence of increased surface temperatures, increased surface and mid-level humidity and broad scale ascent. We suggest that the warmer, more moist conditions at the surface combined with favorable synoptic features lead to increased CAPE over the Gulf Stream, promoting stronger vertical motion and enhanced mixed-phase microphysics, leading to higher lightning flash rates relative to adjacent coastal waters.
AE33A-0186 1340h
Relationship Between Cloud Dynamics and Electrification for a Mid-Latitude Squall Line
During the 2004 TELEX field program, simultaneous electric field soundings, three-dimensional lightning mapping observations, high- resolution dual-Doppler radar data, polarimetric radar data, and environmental soundings were acquired for several mesoscale convective systems, supercell storms, and non-severe thunderstorms. The overall data set was of particularly high quality for a squall line that produced frequent lightning in southern and central Oklahoma on 19 June 2004. A total of five balloon-borne electric field soundings were launched into the leading line of convection and into the trailing stratiform region. Two mobile 5-cm wavelength Doppler radars collected coordinated volume scans every 3 min throughout the period of operations. These data provide near continuous coverage of the flow through the convective system. Furthermore, all operations were well within range of a 10-cm wavelength, polarimetric radar and a three-dimensional lightning mapping array throughout the period. This presentation will emphasize the evolution of the airflow through the convective system and relate that to the structure of electriic fields and frequency of in-cloud and cloud-to-ground lightning.
AE33A-0187 1340h
The Contrast of the El Ni\~{n}o and La Ni\~{n}a Events on the Convection Over East Asia
The 1997-1998 El Ni\~{n}o is one of the strongest El Ni\~{n}o Southern Oscillation (ENSO) events. The 1997-1998 El Ni\~{n}o began to develop in March, 1997 and strengthened rapidly over the next year. After a gradual decline in the intensity of these thermal anomalies in early 1998, the El Ni\~{n}o event abruptly ended during May-June, 1998. After this strongest El Ni\~{n}o event, La Ni\~{n}a event started soon. This La Ni\~{n}a event continued more than a year, and then disappeared in the spring of 2000. {\it Lightning Research Group of Osaka University} (LRG-OU) asserted that the lightning activities are more frequent despite less convective storms during the El Ni\~{n}o period than during La Ni\~{n}a period in East Asia, especially in Indonesia. In this paper, the authors pay their attention to the contrast of the convective activities between El Ni\~{n}o and La Ni\~{n}a periods over the East Asia. The effects of El Ni\~{n}o and La Ni\~{n}a events to the precipitation and lightning activities are investigated. Their possible correlation with some parameters of storms is also considered. The authors studied the convection and lightning activities over the Pacific Ocean and the surrounding Maritime Continent, especially in East Asia, in terms of the Tropical Rainfall Measuring Mission (TRMM) observations. The data of lightning flash and precipitation are obtained from Lightning Imaging Sensor (LIS) and Precipitation Radar (PR) on TRMM, respectively. These PR data include the storm heights, number of observed precipitation, and so on. The increase in number of lightning flashes during El Ni\~{n}o over East Asia has been studied by making a comparison of convective activities between April, 1998, corresponding to El Ni\~{n}o period, and April, 1999, corresponding to El Ni\~{n}o period. The main differences are summarized as follows; 1) During the El Ni\~{n}o/La Ni\~{n}a period the number of lightning flashes increases/decreases. 2) During the El Ni\~{n}o/La Ni\~{n}a period the times of precipitation decreases/increases, but the mean intensity of precipitation is stronger/weaker than Normal period. So the total amount of precipitation is low/high. 3) During the El Ni\~{n}o/La Ni\~{n}a period the storms arise less/more, but the higher storms over 7000m are generated more/less. These results lead to the conclusion that during the La Ni\~{n}a period many storms arise and make a lot of precipitation, but storms do not grow up so high as to make heavy lightning activity. On the contrary, during the El Ni\~{n}o period the storms do not arise so frequently and consequently it does not rain so often. But more storms grow up to high altitude, and those storms increase the lightning flashes. Tough the results in this study are preliminary and further investigation and more evidences are needed, we found the implications of the change in convective precipitation due to El Ni\~{n}o and La Ni\~{n}a events on the lightning activity and their possible correlation with some parameters for storms.
AE33A-0188 1340h
A Composite Evolution of Radar and Electrical Characteristics of Small Thunderstorms over the Florida Everglades
During August 2002, the Altus Cumulus and Electrification Study (ACES) was conducted over south Florida. The ACES project employed an unmanned aerial vehicle (UAV) to fly over thunderstorms and collect storm electrical measurements (including total lightning, electric fields, and electric field changes). The UAV flew at high altitude (~15 km) and its high maneuverability enabled near continuous monitoring of a number of thunderstorms during the project. This study combines observations of thunderstorms in varying stages of their life-cycles to describe the radar and electrical evolution of a composite single cell thunderstorm. National Weather Service WSR88D radar reflectivity data are used to derive storm characteristics including storm mass, volume integrated liquid (VIL), vertical reflectivity distribution, rainflux, and storm height. The radar derived parameters are then compared to the electrical measurements. The composited storm evolution compares favorably with previous studies.
AE33A-0189 1340h
Electric field Change Characteristics of Strong and Weak Optical Pulses from the Altus Cumulus Electrification Study (ACES)
We examine the properties of the electric field changes associated with lightning generated optical pulses recorded as part of the Altus Cumulus Electrification Study (ACES). During ACES, we flew an unmanned aerial vehicle (UAV) around and over thunderstorms near the Western edge of the Florida Everglades. ACES was based out of Key West, Florida and operated during the month of August, 2002. During the nine data flights, we recorded time resolved optical and electrical pulses from the lightning generated in the thunderstorms. We have previously described the properties of lightning generated optical waveforms recorded during ACES and in that study, we found a class of optical pulses with lower amplitudes than what has been previously described. The possible sources of the weak optical pulses ranged from weak lightning near the top of the storms to artifacts due to our close proximity to the storm top (i.e., reflections off of the cloud tops from more distant lightning, or lightning from within the cloud partially obscured by local cloud elements). In this current study, we manually examine the electric field and optical pulse characteristics (risetimes, amplitudes, ratios, delays, etc.) of a subset of the 1172 weak and 3141 strong pulses to see if the waveforms can provide more information as to the source of the weak pulses.
AE33A-0190 1340h
Cloud-to-ground Lightning, Convective Rainfall, and Post-wildfire Flash Floods in Colorado
This project investigates the relationship between hazardous hydrological responses, convective rainfall, and cloud-to-ground (CG) lightning flash parameters. Drainage basins burned by wildfires during the summer of 2002 in western Colorado were identified and the NCDC publication Storm Data provided time and location information for 12 flash floods occurring in those basins. Atmospheric circulation associated with the North American Monsoon (NAM) contributed low level moisture to the Rocky Mountain region during each of the flash floods. A composite analysis of the 12 flash flood episodes revealed a 500hPa equivalent potential temperature (theta-e) ridge extended into Nebraska and southern South Dakota, and theta-e values as high as 338 K at 700hPa over central Colorado. Storm-total CG flashes peaked at 718 flashes. Mean CG flash intensity for the 12 events was 18.2cgf/5-min. Mean rainfall intensity associated with the hydrological events was 5.67mm/5-min, and mean storm-total rainfall was 14.2mm. Continuous CG flash sequences were significantly correlated with both rainfall intensity and total rainfall during post-wildfire flash flood episodes. Total CG flashes were also significantly correlated with rainfall intensity. The mean lag-time from the first recorded CG flash to flash flood report was 145 minutes, and the mean lag from peak CG flash intensity to flash flood report was 35 minutes.
AE33A-0191 1340h
Duration and Extent of Large Electric Fields in a Thunderstorm Anvil After the Last Lightning Flash
In 1999 we obtained a series of three complete soundings of electric field (E) through a large anvil. The soundings, acquired with balloons launched from Langmuir Laboratory in New Mexico, cover a 75-minute period during which the anvil moved northward and dissipated. Each of the balloons ascended through but near the edge of the radar-detectable anvil cloud. The later two soundings were made after the last lightning had occurred in the storm. The second balloon was launched 11 min after and 15 km distant from the closest part of the final lightning flash, and the third balloon was launched 28 min later. The first sounding in the series shows large electric field magnitudes (maximum 65 kV m$^{-1}$) and significant charge densities of both polarities (maximum positive charge density +2.5 nC m$^{-3}$) within and screening the anvil. There was little change in the maximum E values during the intervening 30 min between the first and second balloon flights, although the altitudes and densities of some charge regions were appreciably lower in the second sounding (maximum positive charge density +0.6 nC m$^{-3}$). The final balloon sounding was made through the visible, but optically less dense anvil cloud; maximum radar reflectivity along the track was 12 dBZ. There was essentially no enhanced E or net charge detected in this part of the anvil cloud. In this presentation we compare the location of the radar echo boundaries to the three anvil E soundings made near and through these boundaries. An estimate of the horizontal extent and duration of the charged regions within the anvil can be made from these data. For example, 30 min after the last lightning flash, we estimate that the interior positive charge region covered at least 130 km${^2}$, and 50 min after, it covered less than 50 km${^2}$. We also investigate the evolution of the screening layers and their location relative to the echo boundaries. These analyses should allow us to estimate the extent of potentially hazardous electric fields associated with an anvil after the last lightning flash.
AE33A-0192 1340h
Detailed Lightning Observations With the Compact Lightning Mapping Array at Langmuir Laboratory
New Mexico Tech's three-dimensional Lightning Mapping Array (LMA) is being operated in a compact configuration at Langmuir Laboratory. Eight of the twelve mapping stations are deployed within a 4-km diameter area around the mountain laboratory and the system is being operated in high time-resolution mode, allowing sources to be located in successive 10 microsecond windows. The relatively small size of the array together with the high time resolution will allow us to study close lightning flashes in great detail. The data are augmented with `fast' and `slow' electric field change recordings of the lightning, also beneath the storm. The combination is to be used for detailed studies of lightning discharge processes, examples of which will be presented. 400-microsecond time resolution data from the mapping array are processed in real time and made available on the web at http://lightning.nmt.edu/ll\_lma. This allows us to monitor storms as they develop and to facilitate other research activities such as balloon flights into the storms. An interesting result of having a compact array is that the mapping network locates the sources of strong corona from power lines on the plains around the mountain.
AE33A-0193 1340h
Real Time Display of New Mexico Tech Lighning Mapping Array Data
The New Mexico Tech Lightning Mapping Array (LMA) generates highly detailed three-dimensional images of lightning discharges over an area of about 150~km from the array center. A typical LMA consists of twelve VHF receivers distributed over an area about 80~km in diameter. Each receiver determines the amplitude and time of the highest VHF pulse in each 80~$\mu$s window -- those signals whose amplitudes exceed a noise-floor threshold are recorded to a local hard drive. During an active thunderstorm, data rates for each receiver are as high as 750 KB/s. With increases in computing power and wireless data speed, it is now possible to transmit and process a sufficient volume of data from the VHF receivers that images with adequate detail for real-time uses can be generated. There are currently three LMAs which have real-time displays which are accessible over the Internet. NASA has developed their own display for their North Alabama LMA$^1$, which has been available for the past several years. During the spring and summer of 2004 we developed real-time displays for the NSSL/Oklahoma University LMA$^2$ and for the NMT compact LMA installed at Langmuir Laboratory$^3$. A display for a fourth system, currently being installed at White Sands Missile Range in southern New Mexico, should be on-line by the summer of 2005. The eleven VHF receiving stations of the Oklahoma LMA are connected to a central processing computer over IEEE~802.11b wireless data links. Severn receivers are daisy-chained through one link and four stations are daisy-chained through a second link. Data rates are too high to allow transmission of the full data over the daisy-chained links, so the data transmitted by each station is reduced such that only the highest peak in a 400~$\mu$s window is transmitted. Each station sends its data to the central site in one-minute packets. After data from a sufficient number of stations (a minimum of six) have been received by the processing computer, the data is processed to generate RF source locations. A source location is generate about two to three minutes after the parent event. The location data is sent to the display computer, where plots are generated and posted on a web site. Plots of ten minutes of data and of one hour of data are available. Data are displayed in terms of the density of RF sources per unit area. The compact LMA at Langmuir Lab (see Edens et al, this conference), in the Magdalena Mountains of central New Mexico, has eight stations distributed over an area of about 4~km in diameter, 2~km above the surrounding plains. The array detects noise sources (e.g., corona from power lines) on the plains surrounding the mountains. Because of the compactness of the array, these localized sources appear on the plots as a line extending radially away from the center of the array through the sources. During periods of no lightning activity, the display often shows several radial lines from such localized noise sources. $^1$http://branch.nsstc.nasa.gov/cgi-bin/LMA.pl $^2$http://lightning.nmt.edu/oklma $^3$http://lightning.nmt.edu/ll_lma