National Lightning Detection Network Data Research Articles

ENSO and MJO Modulation of U.S. Cloud-to-Ground Lightning Activity

Abstract Cloud-to-ground (CG) lightning substantially impacts human health and property. However, the relations between U.S. lightning activity and the Madden–Julian oscillation (MJO) and El Niño–Southern Oscillation (ENSO), two predictable drivers of global climate variability, remain uncertain, in part because most lightning datasets have short records that cannot robustly reveal MJO- and ENSO-related patterns. To overcome this limitation, we developed an empirical model of 6-hourly lightning flash count over the contiguous United States (CONUS) using environmental variables (convective available potential energy and precipitation) and National Lightning Detection Network data for 2003–16. This model is shown to reproduce the observed daily and seasonal variability of lightning over most of CONUS. Then, the empirical model was applied to construct a proxy lightning dataset for the period 1979–2021, which was used to investigate the summer MJO–lightning relationship at daily resolution and the winter–spring ENSO–lightning relationship at seasonal resolution. Overall, no robust relationship between MJO phase and lightning patterns was found when seasonality was taken into consideration. El Niño is associated with increased lightning activity over the coastal Southeast United States during early winter, the Southwest during winter through spring, and the Northwest during late spring, whereas La Niña is associated with increased lightning activity over the Tennessee River valley during winter. Significance Statement Cloud-to-ground lightning is dangerous for humans via direct strikes or through triggering wildfires, generating air pollution, etc. How lightning activity can be affected by climate remains unclear, and it is challenging to study their links because the data record for lightning is short. With the available lightning record, we developed a model that relates lightning flash counts over the United States to environmental factors. This model well represents observed fluctuations in daily and seasonal lightning over most of the United States. Because the model only needs environmental information to predict lightning flash counts, we were able to construct a longer record of predicted lightning based on the longer data record of environmental variables. With this dataset, we investigated the links between lightning and climate and found that the state of sea surface temperatures in the tropical Pacific (El Niño–Southern Oscillation) is linked to changes in U.S. lightning patterns in winter and spring.

Assessing the Impact of Lightning NOx Emissions in CMAQ Using Lightning Flash Data from WWLLN over the Contiguous United States

Comparison of lightning flash data from the National Lightning Detection Network (NLDN) and from the World Wide Lightning Location Network (WWLLN) over the contiguous United States (CONUS) for the 2016–2018 period reveals temporally and spatially varying flash rates that would influence lightning NOx (LNOx) production due to known detection efficiency differences especially during summer months over land (versus over ocean). However, the lightning flash density differences between the two networks show persistent seasonal patterns over geographical regions (e.g., land versus ocean). Since the NLDN data are considered to have higher accuracy (lightning detection with >95% efficiency), we developed scaling factors for the WWLLN flash data based on the ratios of WWLLN to NLDN flash data over time (months of year) and space. In this study, sensitivity simulations using the Community Multiscale Air Quality (CMAQ) model are performed utilizing the original data sets (both NLDN and WWLLN) and the scaled WWLLN flash data for LNOx production over the CONUS. The model performance of using the different lightning flash datasets for ambient O3 and NOx mixing ratios that are directly impacted by LNOx emissions and the wet and dry deposition of oxidized nitrogen species that are indirectly impacted by LNOx emissions is assessed based on comparisons with ground-based observations, vertical profile measurements, and satellite products. During summer months, the original WWLLN data produced less LNOx emissions (due to its lower lightning detection efficiency) compared to the NLDN data, which resulted in less improvement in model performance than the simulation using NLDN data as compared to the simulation without any LNOx emissions. However, the scaled WWLLN data produced LNOx estimates and model performance comparable with the NLDN data, suggesting that scaled WWLLN may be used as a substitute for the NLDN data to provide LNOx estimates in air quality models when the NLDN data are not available (e.g., due to prohibitive cost or lack of spatial coverage).

Analysis of Location Errors of the U.S. National Lightning Detection Network Using Lightning Strikes to Towers

AbstractThe location accuracy of the U.S. National Lightning Detection Network (NLDN) has been evaluated using as ground‐truth rocket‐triggered lightning data or video records but only at a few specific locations. In this study, by using the NLDN data for the events attributable to lightning strikes to towers, the location error of the NLDN across the entire contiguous United States was evaluated for the first time. We found that, on average, the NLDN median location error reduced from 198 to 84 m after the 2013 NLDN upgrade. The location error at the periphery of the network is significantly larger than that in its interior. In the coastal regions, there is directional location bias toward the water. Simulation results suggest that the bias is related to the lengthening of field waveform front due to electromagnetic wave propagation over lossy ground coupled with the asymmetrical sensor configuration relative to the strike point (lack of offshore sensors).

Cloud-to-Ground Lightning Flash Density and Thunderstorm Day Distributions over the Contiguous United States Derived from NLDN Measurements: 1993–2018

Abstract This study employs cloud-to-ground (CG) lightning flash data from the U.S. National Lightning Detection Network (NLDN) to examine temporal and spatial distributions of lightning flash and thunderstorm day (TD) occurrences over the contiguous United States from 1993 to 2018. TD distributions are estimated from NLDN CG flashes using 4 thunder audibility approximations: 5 and 10 nautical mile (n mi; 1 n mi = 1.852 km) audibility ranges, and minima of 1 and 2 flashes within the audibility range. The 26-yr period examined is longer than previous studies using NLDN data, and the TD results can be compared directly to climatologies derived from surface weather observations dating back to the late 1890s. Results based on the abundant NLDN data avoid limitations introduced by the coarse horizontal resolution of surface observations inherent in pre-NLDN TD climatologies. Annual mean flash density and annual and monthly mean TD distributions are derived from almost 568 million NLDN CG flashes. A mean annual maximum of more than 16 flashes km−2 is found near Tampa, Florida. The mean annual TD maximum of 113 days (from at least 2 flashes within 10 n mi) occurs in southern Florida. Regions exceeding 70 TDs are found from eastern Texas eastward into Florida, and over the southern Rocky Mountains. Large positive deviations from the mean number of TDs extend from Texas northwestward into Colorado during 2003–07, followed by large negative deviations over the same region during 2008–12. Both deviation patterns are similar to expected summertime precipitation anomaly patterns over the United States during El Niño and La Niña years, respectively.

Antenna structures and cloud‐to‐ground lightning location: 1995–2015

AbstractSpatial analyses of cloud‐to‐ground (CG) lightning occurrence due to a rapid expansion in the number of antenna towers across the United States are explored by gridding 20 years of National Lightning Detection Network data at 500 m spatial resolution. The 99.8% of grid cells with ≥100 CGs were within 1 km of an antenna tower registered with the Federal Communications Commission. Tower height is positively correlated with CG occurrence; towers taller than 400 m above ground level experience a median increase of 150% in CG lightning density compared to a region 2 km to 5 km away. In the northern Great Plains, the cumulative CG lightning density near the tower was around 138% (117%) higher than a region 2 to 5 km away in the September–February (March–August) months. Higher CG frequencies typically also occur in the first full year following new tower construction, creating new lightning hot spots.

Synoptic scale outbreak of self‐initiated upward lightning (SIUL) from tall structures during the central U.S. blizzard of 1–2 February 2011

AbstractA major central U.S. winter cyclone on 1–2 February 2011 produced a band of high winds, up to 75 cm of snow, and numerous reports of thundersnow from Oklahoma into Ontario over a 26 h period. The National Lightning Detection Network (NLDN) recorded 282 flashes comprised of 1153 events which were >96% negative polarity. Hopes of imaging winter sprites associated with energetic positive cloud‐to‐ground events that sometimes accompany such winter storms did not materialize. However, the lack of lightning over the Great Lakes waters, plus media reports of numerous thundersnow events in downtown Chicago, prompted a detailed analysis of the NLDN data. This revealed that >93% of all lightning in the snow band was likely or possibly associated with self‐initiated upward lightning (SIUL) events from a variety of tall, and some not so tall, structures. In addition to 43 events from two Chicago skyscrapers, many shorter structures were involved, including wind turbines (13.1% of the total) and transmission line towers (6.7%). Wind speeds for all events exceeded the 8 m s−1 minimum threshold associated with SIULs in Japanese winter lightning storms. Radar reflectivities at the event locations had a mean of 28 dBZ and were almost always <35 dBZ. While conventional radar displays suggested stratiform precipitation in the thundersnow region, detailed analysis of 3‐D‐gridded NMQ (National Mosaic and Multi‐Sensor Quantitative Precipitation Estimation) radar reflectivity data confirmed elevated embedded cellular convection spanning the −10°C region associated with isentropic lifting above a frontal surface, evidence of noninductive charge generation sufficient to allow upward leader initiation from tall objects.

Observations of the 14 July 2011 Fort Collins Hailstorm: Implications for WSR-88D-Based Hail Detection and Warnings

Abstract The issuance of timely warnings for the occurrence of severe-class hail (hailstone diameters of 2.5 cm or larger) remains an ongoing challenge for operational forecasters. This study examines the application of two remotely sensed data sources between 0100 and 0400 UTC 14 July 2011 when pulse-type severe thunderstorms occurred in the jurisdiction of the Denver/Boulder National Weather Service (NWS) Forecast Office in Colorado. First, a developing hailstorm was jointly observed by the dual-polarization Colorado State University–University of Chicago–Illinois State Water Survey (CSU–CHILL) research radar and by the operational, single-polarization NWS radar at Denver/Front Range (KFTG). During the time period leading up to the issuance of the initial severe thunderstorm warning, the dual-polarization radar data near the 0 °C altitude contained a positive differential reflectivity ZDR column (indicating a strong updraft lofting supercooled raindrops above the freezing level). Correlation coefficient ρHV reductions to ~0.93, probably due to the presence of growing hailstones, were observed above the freezing level in portions of the developing >55-dBZ echo core. Second, data from the National Lightning Detection Network (NLDN), including the locations and polarity of cloud-to-ground (CG) discharges produced by several of the evening’s storms, were processed. Some association was found between the prevalence of positive CGs and storms that produced severe hail. The analyses indicate that the use of the dual-polarization data provided by the upgraded Weather Surveillance Radar-1988 Doppler (WSR-88D), in combination with the NLDN data stream, can assist operational forecasters in the real-time identification of thunderstorms that pose a severe hail threat.

Upward lightning observations from towers in Rapid City, South Dakota and comparison with National Lightning Detection Network data, 2004–2010

We report on upward lightning observations from ten tall towers (91–191 m) in Rapid City, South Dakota, USA and compare with National Lightning Detection Network (NLDN) data. A total of 81 upward flashes were observed from 2004–2010 using GPS time‐stamped optical sensors, and in all but one case, visible flash activity preceded the development of the upward leaders. Time‐correlated analysis showed that the NLDN recorded an event within 50 km of towers and within 500 ms prior to upward leader development from the tower(s) for 83% (67/81) of the upward flashes. A preceding positive cloud‐to‐ground stroke (+CG) was detected in 57% (46/81) of the cases, and a preceding positive intracloud flash (+IC) in 23% (19/81) of the cases. However, 8 of the 19 NLDN‐indicated +IC events were actually +CG strokes based on optical observations. Preceding negative intracloud flashes (−IC) were recorded for 2% (2/81) of the cases. Analysis also showed that for 44% (36/81) of the upward flashes, the NLDN reported subsequent negative cloud‐to‐ground (−CG) strokes and/or −IC events at one or more tower locations. Of the 151 subsequent events, 70% (105/151) were −CG reports and 30% (46/151) were listed as −IC events. The geometric mean/median location accuracy and peak current for subsequent events were 194 m/206 m and −12.9 kA/−12.4 kA respectively. These correlated observations suggest that a majority of the upward lightning flashes were triggered by a preceding flash with the dominant triggering type being the +CG flash.

The effect of lightning NO<sub>x</sub> production on surface ozone in the continental United States

Abstract. Lightning NOx emissions calculated using the US National Lightning Detection Network data were found to account for 30% of the total NOx emissions for July–August 2004, a period chosen both for having higher lightning NOx production and high ozone levels, thus maximizing the likelihood that such emissions could impact peak ozone levels. Including such emissions led to modest, but sometimes significant increases in simulated surface ozone when using the Community Multi-scale Air Quality Model (CMAQ). Three model simulations were performed, two with the addition of lightning NOx emissions, and one without. Domain-wide daily maximum 8-h ozone changes due to lightning NOx were less than 2 ppbv in 71% of the cases with a maximum of 10 ppbv; whereas the difference in 1-h ozone was less than 2 ppbv in 77% of the cases with a maximum of 6 ppbv. Daily maximum 1-h and 8-h ozone for grids containing O3 monitoring stations changed slightly, with more than 43% of the cases differing less than 2 ppbv. The greatest differences were 42 ppbv for both 1-h and 8-h O3, though these tended to be on days of lower ozone. Lightning impacts on the season-wide maximum 1-h and 8-h averaged ozone decreased starting from the 1st to 4th highest values (an average of 4th highest, 8-h values is used for attainment demonstration in the US). Background ozone values from the y-intercept of O3 versus NOz curve were 42.2 and 43.9 ppbv for simulations without and with lightning emissions, respectively. Results from both simulations with lightning NOx suggest that while North American lightning production of NOx can lead to significant local impacts on a few occasions, they will have a relatively small impact on typical maximum levels and determination of Policy Relevant Background levels.

Cloud-to-Ground Lightning Flash Parameters Associated with Heavy Rainfall Alarms in the Denver, Colorado, Urban Drainage and Flood Control District ALERT Network

Abstract Rainfall data from the Denver, Colorado, Urban Drainage and Flood Control District Automated Local Evaluation in Real Time (ALERT) network were used to identify heavy rainfall alarms for the period 1999–2003. Twenty-nine heavy rainfall-rate alarms were identified. Cloud-to-ground (CG) lightning flash data from the National Lightning Detection Network (NLDN) were analyzed for the 90 min prior to each heavy rainfall alarm. Spatial patterns from NLDN data were extracted using a point-polygon topology developed with basic Geographic Information System procedures. The information extracted from the polygons was used to calculated summary statistics for rainfall rates, CG flash rates, and CG flash duration. Heavy rainfall episodes were divided into two groups based on latitude, longitude, and elevation. Heavy rainfall episodes in the higher elevations of the study area produced an average of 29 mm of rainfall per episode and 1095 CG flashes in the 90 min prior to the rainfall-rate alarm. Only five polygons, all closely proximal to the alarm sites, produced significant CG flash rates prior to the rainfall alarms, and areas with CG flash durations greater than 25 min were clustered near the rainfall-rate alarm sites. In the second group (the lower elevation stations) the mean event produced a total of 33 mm of rainfall and 1182 CG flashes during the 90 min prior to the rainfall alarm. Four polygons saw consistent CG flash rates in the 90 min prior to the heavy rainfall alarms and CG flash duration was at its greatest in areas just west of the ALERT stations.

Determining the striking distance of lightning through its relationship to leader potential

A new method for estimating lightning leader potential in negative stepped leaders from electric field changes during the first return stroke process in cloud‐to‐ground flashes is applied for determining the striking distance to a ground structure. This method may serve as an alternative to calculating the striking distance from formulas that relate it to the peak return stroke current. Presented are the characteristics of downward leaders (potential, vertical extent, and the length of the final step) and the charges transferred during return strokes; these were obtained from electric field measurements during a thunderstorm in Florida using the line charge model. The final step lengths of the leaders are compared with striking distances calculated using the peak current values of return strokes taken from National Lightning Detection Network data for the same flashes.

Determination of source thunderstorms for VHF emissions observed by the FORTE satellite

This paper explains a method of locating groups of satellite‐observed VHF signals from lightning that emanate from isolated storm regions. LF/VLF signals recorded and located by the National Lightning Detection Network (NLDN) are often accompanied by broadband VHF emissions of sufficient intensity to trigger satellite‐based RF receivers. It has previously been shown that events recorded by the FORTE satellite and the NLDN, which have occurrence‐time differences of 0.3 ms or less, can be assumed to originate at the same approximate geolocation with 97.5% reliability. For such event pairs, the VHF events recorded by FORTE were assigned the NLDN locations. We use graphical analysis to identify groups of satellite data that can be assumed to originate from the same storms as the located events. To accomplish this, a value of total electron content (TEC) is derived from the frequency dispersion of each VHF signal recorded by FORTE. Scatterplots of TEC versus time often reveal curved clusters of satellite data. The variation of TEC with satellite elevation angle, and time, is consistent with a model in which the length of a signal's required propagation path through the ionosphere changes with satellite and storm position. We use the values of TEC for the located satellite data, a simple model of the ionosphere, and NLDN data to help us identify TEC clusters that can be assumed to originate from relatively compact, and unambiguous, geolocations. For the time period of April through September of 1998 we geolocate 65 groups of satellite data records. A total of 6131 satellite data records, which were not previously located, have been assigned the median NLDN coordinates of their corresponding storm.

FORTE radio‐frequency observations of lightning strokes detected by the National Lightning Detection Network

This work compares simultaneous observations of lightning from two complementary systems. FORTE is a low‐Earth‐orbit satellite carrying radiowave and optical instruments for the study of lightning. The radio receivers aboard FORTE observe very high frequency (VHF) emissions from the air‐breakdown process preceding (and sometimes accompanying) a lightning current. The National Lightning Detection Network (NLDN) is a ground‐based array of sensors in the contiguous United States observing the low‐frequency (LF) and very low frequency (VLF) radiation from vertical currents. Prior to the launch of FORTE in 1997, essentially no work had been done on the statistical correlations between (1) ground‐based LF/VLF and (2) spaced‐based VHF remote sensing of lightning. During a 6‐month campaign in April‐September 1998, FORTE took most of its triggered VHF data over and near the contiguous United States, and NLDN data were specially postprocessed in a loosened‐criterion mode providing enhanced detection range beyond the coastline and borders of the array itself. The time history of reported events from the two systems was compared, and event pairs (each pair containing one event from FORTE, the other from NLDN) which were candidate correlations (closer than 200 ms from each other) were scrutinized to determine whether the members of a pair actually came from the same discharge process. We have found that there is a statistically significant correlation, for a subset of FORTE events. This correlation is most likely to occur for intracloud and less likely to occur for cloud‐to‐ground discharges. The correlated VHF and NLDN events tend to occur within ±30 μs of each other, after correction for the propagation of the VHF signal to FORTE from the NLDN‐geolocated stroke location. Most correlations outside of ±30 μs turn out to be merely a statistical accident. The NLDN‐furnished geolocation allows the correlated FORTE‐detected VHF pulses to be better interpreted. In particular, we can deduce, from the lag of the VHF ground‐reflection echo, the height of the VHF emission region in the storm.

Performance evaluation of the U.S. National Lightning Detection Network in eastern New York: 2. Location accuracy

We have evaluated the location accuracy of the U.S. National Lightning Detection Network (NLDN) via comparison of individual NLDN stroke locations with their corresponding stroke terminations as determined from multiple‐site video recordings of local cloud‐to‐ground lightning flashes. The video records used in this analysis were acquired in the vicinity of Albany, New York, during the summers of 1994 and 1995, seasons corresponding to the initial and final phases of the upgrade of the NLDN to the improved accuracy from combined technology (IMPACT) configuration. For 1994 a total of 751 strokes were located in common between the video and NLDN data sets, yielding median and mean values of the NLDN‐video location separation of 2.21 km and 3.74 km, respectively. Constraint of the 1994 data to the best video locations (those with a semimajor axis value of the 50% error ellipse less than 500 m) yielded a subset of 53 strokes with median and mean values of the NLDN‐video location separation of 2.61 km and 4.74 km, respectively. For 1995 a total of 219 strokes were located in common, yielding median and mean values of the NLDN‐video location separation of 442 m and 865 m, respectively. Constraint of the 1995 data to that year's best video locations (those with a semimajor axis value of the 50% error ellipse less than 200 m) yielded a subset of 79 common strokes with median and mean values of the NLDN‐video location separation of 435 m and 625 m, respectively. The IMPACT upgrade of the NLDN apparently has resulted in about a fivefold increase in location accuracy. A special subset of 11 strokes occurred in 1995; these were located by the NLDN and observed to terminate on local structures of accurately known location. For these 11 strokes, the median and mean values of the NLDN location error are 518 m and 484 m, respectively, a result quite consistent with the overall NLDN‐video location comparison and the claimed 500‐m median location accuracy for the upgraded network. Indeed, we argue that removal of an apparent residual bias in the NLDN locations ultimately could yield a median location accuracy in this region approaching 250 m. Finally, we find that the NLDN location error is a function of the estimated stroke peak current, with the weakest peak currents having the poorest locations for both years. For the 1995 data, strokes with peak currents of intermediate magnitude (20–30 kA) were the most accurately located.

A reexamination of the peak current calibration of the National Lightning Detection Network

The peak current calibration of the National Lightning Detection Network (NLDN) recently reported by Orville (1991) has been reexamined with a roughly threefold larger data set of 57 directly measured stroke peak currents, Ipeak (kiloamperes), and their corresponding NLDN mean normalized magnetic signal strengths, (LLP units). These 57 lightning strokes originated in 36 triggered lightning flashes initiated at the Kennedy Space Center (KSC), Florida, during the research campaigns of 1985–1991. Identification of corresponding Ipeak and measurements was verified through accurate coincidence in absolute time of the two independent data sets. The data (with one point excluded as an outlier) is apparently linearly related with a correlation coefficient of 0.881, consistent with that predicted by application of the transmission line model of the lightning return stroke. The regression equation for prediction of Ipeak from NLDN measurements is where the slope is expressed in kiloampere/(LLP units). Examination of the overall data set for the possible influence of two different models of signal strength attenuation with distance, D, (power law, Dβ, and exponential, [expαDD]−1) indicates negligible sensitivity to the proposed variations; other larger error sources likely mask the true attenuation effect. Twelve flashes were detected with four or more direction finders; a power law fit to the direction finder signal strength variation with distance of these individual flashes yields a mean β value of −1.09. Examination of the overall data set for the possible effect of a nonlinear relation between the source stroke peak current and return stroke propagation speed indicates no obvious influence. The 95% confidence bounds for peak current prediction in the mean suggest a percent uncertainty of 10–15% for peak currents between 15 and 60 kA. Similar accuracy is expected for peak currents greater than 60 kA provided care is taken to minimize possible nonlinear amplification effects hi the NLDN data. Below 15 kA, the percent uncertainty rapidly increases suggesting that the inferred distribution of peak current values less than this limit may be quite unreliable.