Pacific Lightning Detection Network

Introduction

PacNet is a project to continuously monitor convective storms over the Pacific ocean. PacNet is based on a network of four lightning location sensors in the central north Pacific. The detectors have been installed in Dutch Harbor, Kwajalein Atoll, Lihue and Kona (Fig. 7). Fifth site is planned in Kiritimati (Christmas Island).

The goal of the network is to monitor the location of electrical activity associated with thunderstorms. These thunderstorms represent a threat to ocean shipping and airborne carriers and are beyond the range of weather radar. The network also promises to give insight into the intensification of tropical cyclones and flash flood events. Additionally the lightning detectors provide ground truth for a NASA satellite based instrument and a data stream for ingest into global weather prediction models.

Scientific Objectives

The wave guide between the Ionosphere and the Earth's surface allows VLF noise generated by lightning between 5 and 15 kHz to propagate over very long distances. Thunderstorms are widespread over the Pacific Ocean, especially at low latitudes, along the Pacific Rim and over East Asia (Fig. 1). Lightning detection has great value for real-time storm warning, tracking, and (recently) prediction. In remote regions where conventional data sources are not available, tracking of thunderstorms and rapid intensification of cyclones are important challenges in weather prediction for civilian and military purposes. A lack of real-time weather data from the area surrounding Naval aircraft carrier operations and civilian airports on islands presents an important potential application for PacNet. Because of its long range detection capability, PacNet will provide inexpensive, accurate monitoring of convective storms that represent a threat to ocean shipping and airborne carriers and are beyond the range of weather radars. In Hawaii and the surrounding Pacific region there is currently an unmet requirement for long-distance real-time storm tracking.

Fig. 1. An annual summary of lightning strokes
observed from the lightning Imaging Sensor (LIS)
on TRMM. Sferics data collected approximately
twice daily in 90s intervals during 1998. (Christian, 1999)
(click on image to see full size)

Moreover, sferics data will promote fundamental understanding of convective weather systems and the hydrological cycle over the Pacific Ocean basin. The sferics data will be used to support investigations at the University of Hawaii of (i) tropical and subtropical cyclogenesis, (ii) cyclone track forecasting, and (iii) forecasting flash-flood events over the Pacific.

Assimilation into Numerical Models

The resulting data on lightning stroke frequency and location will be used in conjunction with precipitable water data from ground-based GPS receivers and satellite data to provide guidance for real-time storm tracking and warnings over the data-sparse Pacific Ocean. Recent research has shown that lightning frequency and convective rainfall rates are correlated (Fig. 2). Rainfall patterns derived from sferics have been successfully assimilated into numerical weather prediction models for improved weather forecasts (eg. Jones and Macpherson, 1997 and Alexander et al., 1999). Processed data from PacNet will be assimilated into MM5 , run operationally in Hawaii by the principal investigator's group.

Fig. 2. Relationship between Sferics rate and convective rainfall rates. Convective rainfall, R, from TRMM Microwave imager GPROF algorithm averaged over 0.5 x 0.5 degree areas accumulated within 15 minutes of the TRMM overpass time. The points were obtained from histogram matching and the least squares fitted curve is R = 5.87*S 0.46. (Chang et al., 2001)
(Click on image to see full size)

Low orbiting satellites that carry microwave radiometers such as Tropical Rainfall Measuring Mission (TRMM) with its TRMM Microwave Imager (TMI) and the Defense Meteorological Satellite Program (DMSP) satellites carrying the Special Sensor Microwave/Imager (SSM/I) provide intermittent glimpses of convective precipitation. Unfortunately, they do not permit the evolution of convective weather systems to be monitored continuously from space.

Convective precipitation retrievals can be regressed against coincident lightning rate data (as in Fig. 2). Such regressions can then be used to tune continuously inferred lightning data to fill the temporal gaps between overpasses of satellites carrying microwave radiometers. An example of a rainfall distribution retrieved from such "tuned" sferics data is shown in Fig. 3b. The inferred rainfall distribution can be compared to the rainfall (Fig. 3a). The threat score for rainfall rates greater than 1 mm/h in this case was 0.72.distribution derived from coincident TMI data shown in Fig. 3a. Note that the TMI coverage is confined between the parallel lines that define the swath width and the duration of observation is only 90 sec. A regression between the rainfall from the TMI and the STARNET-1 data of Figs. 3a and 3b is shown in Fig. 3c. Lightning has been used to provide bogus latent heating that can be assimilated into mesoscale models (Jones and Macpherson, 1997 and Alexander et al., 1999).

Fig. 3. a) Convective rainfall rates inferred from a TMI orbit that passed over part of the storm at 2010 UTC. b) Convective rainfall derived from empirically tuned STARNET-1 sferics data observed during a 15 minute interval commencing at 2000UTC on Feb. 2, 1998. c) Regression of the rainfall rates in 0.5 x 0.5 grid boxes in Figs 3a) and 3b). (Chang et al., 2001).

Fig. 4 shows that assimilating rainfall data from sferics improved forecasts of precipitation and wind shear more than six hours in advance. That information could mitigate aviation and marine hazards. Intermittent rainfall inferred from data provided by AMSR on the EOS-AQUA platform and from the TMI on TRMM will benefit from complementary continuous sferics data.

Fig. 4. Precipitation field at 0900 UTC on 13 March 1993 from MM5
simulations with progressive data assimilation and from the WSR-88D
radar network. (Alexander et al., 1999)
(Click on image to see full size)

A Global Precipitation Mission (GPM) consisting of six to eight polar orbiting satellites to be launched by NASA and partners in ~2006 will provide more comprehensive microwave observations of rainfall. Once installed, data from PacNet can fill the temporal gap between the end of TRMM and the beginning of the GPM. The sferics data also complement time series of precipitable water obtained from co-located GPS receivers at each of the sites. The availability of a significant data archive of continuous long-range lightning observations will also enable meteorologists to acquire experience with lightning data in preparation for using data from a Lightning Mapping Sensor (LMS) over the western hemisphere (Christian et al., 1989).

Former experiences

Continuous lightning data have been available from the National Lightning Detection Network (NLDN) over the continental U.S. Since extensive meteorological information is already available, the impact of and interest in these lightning data has been limited in the meteorological community. Appreciation of the significance of lightning data has also been mixed. For example, Petersen and Rutledge (1998) discuss the limitations of using lightning rates from NLDN receivers to estimate rainfall rates. They used cloud-to-ground lightning stoke data and rainfall from many sources that included mixtures of stratiform and convective rain, and showed large scatter in the lightning-rainfall proportionality. A Long-Range Lightning Detection Network (LRLDN) (Cramer and Cummins, 1999) using modified NLDN software has provided experimental data during the past two years. Unfortunately, LRLDN placed lightning in regions where no clouds were observed in the IR and the sensitivity is strongly modulated by the diurnal variation of propagation in the Earth-Ionosphere wave guide. However, the PacNet receivers appear to be less affected by those problems. The PacNet sferics network consists of state of the art VLF receivers and software. Unlike the older generation NLDN receivers, which have intrinsic range and detection efficiency limitations, VLF signal propagation over water is less attenuated than over land. Research indicates (Chang et al., 2001) that intermittent tuning of the lightning-rainfall relationship, using convective rainfall derived from coincident space-borne microwave radiometer data in the area will produce good results over the relatively uniform atmospheric environment over the Pacific Ocean (e.g., Fig. 5).

Aviation

Air travel routes over the Pacific Ocean, called tracks, are mapped out before a flight leaves based on weather patterns at the time. The pilot stays on the planned track during an eight-to-ten hour flight even though developing weather patterns may change the optimal route. The pilot uses short-range radar to avoid turbulence in the immediate vicinity, but lacks information about problems in the upcoming part of the flight plan. Fuel economy, passenger safety and comfort could be increased by timely identification of turbulent weather systems. Although GOES and GMS provide a continuous stream of IR imagery, convective activity is often obscured by cirrus anvils (Fig.5). Some of the cirrus covered clouds in Fig. 5 pose little hazard to aircraft operations whereas cold clouds that produce lightning indicate significant updrafts and increased potential of wind shear. Using the lightning data has proved to improve aviation oceanic convective forecasting (Nierow and Showalter, 2000).

Fig. 5. Lightning data from International Long-Range Network (ILRN)
on July 29th 2002, 0654-0954 UTC. Red dots are data from the last hour
and yellow dots from the previous two hours. Lightning observations
are overlaid on infrared imagery obtained from geostationary satellites
(0900 UTC). (Click on image to see full size)

More images (password protected)

Future of the network

Several foreign governments (Japan, Korea, and Australia) have expressed interest in participating in PacNet by deploying additional sferics sensors and sharing data. Therefore, it is anticipated that PacNet will act as a catalyst for eventual Pacific-wide coverage by an international network of sensors. Lightning sensors represent a low-maintenance technology and a manufacturer has expressed interest in providing for ongoing maintenance of the network in exchange for data access. Additionally, funds for to support continued research and future expansion of PacNet will be sought from federal agencies (e.g., NOAA, DARPA, NASA, and ONR) and commercial sources (sensor manufacturer, utility companies).

Instrumentation

The wave guide between the Ionosphere and the Earth's surface allows VLF (Very Low Frequency) noise generated by lightning between 5 and 15 kHz to propagate over very long distances. The lightning detectors (Fig. 6), constructed by Vaisala-GAI Inc. (Global Atmospherics Inc.), have a hybrid design which includes some of the advantages of accuracy at close range of traditional broad-band detectors and the advantages of long range associated with VLF detectors.

IMPACT ESP CLOUD AND CLOUD TO GROUND LIGHTNING SENSOR
Complete sensor dimensions:
Height: 168 cm / 66 in
Width: 34.3 cm / 13.5 in
Depth: 34.3 cm / 13.5 in
Weight: 36.2 kg / 80.5 lbs

Electrical power requirements:
100-230 VAC 47-63 Hz
50 W max. power consumption.

Data communication rate: Able to process and transmit 50 discharges per second at 9600 baud (bits/s) maximum throughput.

Ground mount: The recommended method of pad construction for a ground mount uses a round concrete casting form with a diameter of 61 cm (2 ft). The minimum thickness of the concrete pad should be 61 cm (2 ft).

Direct burial communication cable between sensor and modem/other communication equipment can be up to 4500 m (15000 ft) long. Within a 30-100 m radius, power and communication lines must be buried along a straight line directed toward the sensor mounting pad. Sensor should be located in the center of an open area without any major obstacles in the near distance. Under certain conditions, the sensor can be mounted on a flat roof.

Fig. 6. IMPACT ESP sensor on Lihue airport (Kauai, Hawaii)

An online brochure provides an overview of the detector.

Sensor locations

Table 1. Sensor locations from north to south

(green indicates installed, operational sensor, yellow indicates planned site)

Fig. 7. PacNet sensor sites (red triangles) and potential future sites (open triangles).
Blue and green triangles show some contributing sites in Japan and North America.
(Click on image to see full size)

References

G. D. Alexander et al. 1999: The effect of assimilating rain rates derived from satellites and lightning on forecasts of the 1993 superstorm, Mon. Wea. Rev, 127, 1433-1457.

Chang, Dong-Eon, James A. Weinman, Carlos A. Morales, William S. Olson, 2001: The Effect of Spaceborne Microwave and Ground-Based Continuous Lightning Measurements on Forecasts of the 1998 Groundhog Day Storm. Monthly Weather Review: Vol. 129, No. 8, pp. 1809-1833.

Christian, H.J., 1999: Optical detection of lightning from space, 11th International Conf. on Atm. Electricity, NASA-CP 1999-209261, 715-718.

Christian, H.J., R.J. Blakeslee, S.J. Goodman, 1989: The detection of lightning from geostationary orbit, J.Geophys. Res., 94,13329-13337)

Jones, C.D. and B. Macpherson 1997: Sensitivity of the limited area model to the assimilation of precipitation estimates derived from lightning data , UKMO Forecasting Research Technical Report 212, 11 pp.

Nierow, A. and Showalter R.C., 2000: An Evaluation of using lightning data to improve aviation oceanic feorecasting for the Gulf of Mexico. APIS Severestorm, Sept. 2000.

Petersen, W.A., and S.A. Rutledge, 1998: On the relationship between cloud-to-ground lightning and surface rainfall. J. Geophys. Res., 103, 14025-14040