Use your browser's "find" feature to search for words of interest.

High altitude air breakdown, manifested as "red sprites," is reported in close association with negative cloud-to-ground lightning (-CG) on at least two occasions above an unusual storm on August 29, 1998. Data from high speed photometry, low-light-level video, and receivers of lightning electromagnetic signatures in the frequency range 10 Hz to 20 kHz are used to establish the association and indicate that the causative -CG discharges effected unusually large vertical charge moment changes (Delta M-Qv) of up to 1550 C.km in 5 ms. The existence of sprites caused by -CG's, rather than the regularly associated +CG's, has immediate implications for sprite models and observations.

We measured the charge and size of precipitation particles with an instrumented free balloon in a convective mountain thunderstorm over Langmuir Laboratory in central New Mexico. Using an instrument that measured precipitation charge from 2 to 220 pC and equivalent particle diameters from 0.6 to 3.8 mm, we deduced that (1) the charge of the main positive charge region is carried by cloud particles, the charge of the main negative charge region is carried by a mixture of negative charge bearing cloud particles and precipitation particles, and the charge of the lower positive charge region is carried almost entirely by precipitation; (2) there was a mixture of both polarities of precipitation charge at nearly all altitudes; (3) there was no relationship between precipitation charge and size; (4) our charge data appear to support the noninductive ice-ice collisional charging mechanism; and (5) the sign of charge carried by the precipitation reverses with ambient temperature. In comparing these data to previous measurements that were collected in New Mexico mountain thunderstorms with different instrumentation, we found that in most ways our precipitation charge data are similar to the previous measurements. In comparing these data with data that we previously collected in the trailing stratiform regions of mesoscale convective systems, we found that there are substantial differences in the precipitation charge data between these different cloud types.

We have designed a new instrument to measure the current flowing along balloon rigging line during flights through thunderstorms. This instrument was tested in a high voltage facility and used to collect line current data during one balloon flight into a thunderstorm. Using these data, worst-case calculations are made; as such, we claim that they are the upper limits of any alteration (to the measured electric field or particle charge) that may occur, and the real number is likely much less. It is postulated the rigging-line current could have two separate effects on the measured electric field: (1) reduction of the field due to emission of corona ions, and (2) enhancement of the field due to the insertion of a long thin `conductor.' Even with current as high as 1 $/mu$A (the largest measured was around 50-100 nA), these two effects were found to be about -1% and +1%, respectively. Also, the calculated worst-case alteration to charged precipitation measurements is about 0.1 pC. Thus, with proper efforts to make the rigging line as poor a conductor as possible, it seems that we are justified in stating that these effects are negligible.

A new balloon-borne instrument, created by the authors and referred to as the q-d instrument, that measures the charge q and size d of precipitation particles is discussed. The instrument measures charge with an induction cylinder, size with an optical sensor, and fall speed by the time difference between the two. A second induction cylinder at the top serves as the entry point and detects precipitation that splashes off the entry. In this way, particles contaminated by splashing are removed from the data. It is capable of measuring particle sizes ranging from 0.8 to 8.0 mm in diameter and charges ranging from ± 4 to ± 400 pC. Since the size is measured optically, one can detect uncharged particles and measure their size. The q-d instrument does not show evidence of corona at its extremities until the electric field is as large as 100 kV/m at 700 mb.

Analysis of field observations has yielded the conclusion that the Hallett-Mossop process (H-M) of secondary ice production plays a major role in the glaciation of summertime cumulus clouds over New Mexico. Other studies have revealed that these clouds possess a characteristic multi-thermal structure.

In an effort to quantify more fully the role of H-M in such clouds, and to establish which of the salient dynamical and microphysical parameters play important roles in the glaciation process, a model of ice-particle growth and splinter production in a simple multi-thermal framework is developed. The characteristics of the model are prescribed with values that are based on the above-mentioned field studies. The model cloud possesses four distinct regions: the main updraught, a quiescent (debris) region, the cloud top, and a downdraught region.

The trajectories of all primary ice particles introducted into the cloud at t=0, together with those created as a consequence of the operation of H-M, are followed as they grow and are transported around the cloud. The sensitivities of these trajectories and a multiplication factor f to variations in parameteres such as updraught speed, liquid-water content, L, thermal depth, inter-thermal interval, and downdraught characteristics are examined.

These tests reveal that f is particularly sensitive to the values of L in the distinct regions of the cloud. Basically, combinations of parameter values which produce rapid growth of graupel pellets, large numbers of thermals, and efficient transport between cloud top and the Hallett-Mossop temperature band yield the most rapid ice-particle multiplication.

Observations made in 1987 with the NCAR King Air aircraft and in 1993 with the New Mexico Institute of Mining and Technology dual-polarization radar have revealed the presence of supercooled raindrops in some New Mexico summertime cumulus clouds. In the case of the radar data, the evidence for the supercoled drops came from a column of enhanced Z_DR that extended well above the $0 /deg C$ level. The in situ data indicated that the supercooled raindrops were observed when cloud base was warmer than about $7 /deg C$ and the depth of the cloud was greater than about 2.5 km.

An experiment, involving the National Center for Atmospheric Research's King Air aeroplane, was conducted in order to measure the midrophysical properties of New Mexican summertime cumulus clouds. Since the clouds formed and developed essentially in place, over the mountains, it was possible to make multiple penetrations through a single cloud, thereby observing a significant fraction of the cloud's life cycle. In this paper, the questions of primary- and secondary-ice production, and the development of precipitation particles, are addressed.

Primary-ice nucleation was found to occur when the temperature within the cloud reached a value of between -10 and -12\deg C irrespective of whether this was in the updraught or downdraught. Drops with diameters of about 0.5 mm were often observed in concentrations of about 10 L^-1 before the formation of ice, which suggests a nucleation mechanism involving large drops. The maximum concentrations of ice particles observed in these clouds (up to about 1300 L^-1) are much greater than typical concentrations of ice particles that can be attributed to primary-ice nucleation. Evidence suggests that the most likely explanation is the Hallett-Mossop process of secondary-ice-crystal production.

Ice particles generally were first observed in the downdraughts. The development of precipitation is often thought to occur via downdraught transport, followed by sedimentation or mixing of ice particles into fresh, liquid-laden turrets. The multi-thermal nature of the cloud is considered to be central to this process.

Presented here are the results from a field study concerning the measurement of O-3 and N2O produced by corona discharge from a grounded metal tube with a sharp point during thunderstorms. We observed as much as a tenfold increase in the O-3 concentration as well as a four percent increase in the N2O concentration as a result of corona discharge. The concentrations of both O-3 and N2O returned to the original ambient levels after the thunderstorm ended.

Brandvold, D.K., P. Martinez, and C. Matlock, Method for the determination of mercury in very small solid samples, Anal. Instr., 21, 63--67, 1993.

An alternative method for the determination of mercury (Hg) in solid samples is presented in this article. Gaseous Hg is evolved when a solid sample is heated above 500-degrees-C in the presence of oxygen. The evolved Hg is collected on silver wool and is analyzed using cold vapor methods. We have found greater than 95% Hg recovery with various sediment and ash standards.

Electric field measurements were made by a sailplane inside thunderstorms near Langmuir Laboratory and the Magdalena Mountains in central New Mexico. The continuity of these measurements in-cloud allows us to deduce electric field growth rates in six cases ranging from initial electrification through the production of lightning. The electric field data are combined with radar reflectivity to estimate charge locations and magnitudes. In each case, the dominant charge was negative, associated with a local reflectivity core, and demonstrated a cellular structure centered at altitudes ranging from 5.5 to 7.3 km mean sea level ($-3$ to $-15\C$). Three distinct phases characterize the electric field measurements during initial electrification: an early stage of slowly increasing fields that largely could be explained by the sailplane motion toward the estimated charges, followed by a rapid exponential growth period, and then a plateau in electric field growth rate (at about 300 V/m*s) after the rapid growth period. These six cases extend earlier studies to further establish the character and magnitude of electric field growth during the early stages of thunderstorm electrification.

Lightning flashes lowering positive charge to earth were studied for two diverse systems: the winter thunderstorms of the Hokuriko coast of Japan, and the summer thunderstorms at Kennedy Space Center in Florida. In both regions, a network of electric field-change instruments was used in Japan and ten stations in Florida. The methods used for measutement and analysis are given in Krehbiel et al. [1979].

This paper is the third in a three-part series in which a three-dimensional numerical cloud model is used to simulate cumulus congestus clouds. The authors conduct a detailed parcel trajectory and conserved variable analysis of the modeled clouds, with the principal goal of understanding the mechanisms associated with entrainment and detrainment.

At any point in their lifetime each of the modeled clouds contains multiple thermals that become detached from the boundary layer as they ascend. Undilute regions of subcloud air occur within the simulated clouds at all levels up to the cloud top. In the upper portion of the clouds, such air is found within small (compared with the overall width of the cloud) thermals that are continually eroding yet vigorously ascending. Such thermals are responsible for most of the entrainment and detrainment. Environmental air entrained by ascending thermals is shed in the wake of the thermal, which contains dilute cloud-base air moving at low velocities. There is no evidence for thermals ascending through the remnants of their predecessors as a favored means for new cloud growth. The source of entrained air within both updrafts and downdrafts is typically a few hundred meters above the observation level (although there is a tendency for updrafts at the highest levels to entrain air from just below that level).

Undilute cloud turrets tended to overshoot their level of neutral buoyancy, by a considerable distance. Condenstate loading triggers the collapse of individual turrets, with additional reductions in buoyancy resulting from the evaporative cooling due to entrainment as well as the transport of entrained environmental air upward. Strong, narrow downdrafts develop along the top and edges of overshooting turrets. These downdrafts are often marginally saturated (which would be the most dense mixture of two air masses) and are composed of a mixture of cloud-base and cloud-top air. They descend to mid levels within the modeled clouds before being detrained laterally.

This paper is the second in a three-part series in which a three-dimensional numerical cloud model is used to simulate cumulus congestus clouds at high resolution in an effort to better understand the mechanisms associated with entrainment and detrainment. The prescribed environment is that associated with nonprecipitating summertime New Mexican cumulus clouds that formed on consecutive days. Using budgets of mass and moisture, the effects of the clouds and their environment are examined here with an emphasis on understanding the life cycle of the clouds and the production of narrow detrainment layers aloft. Results are compared with measurements obtained in similar environments.

The mass flux profiles indicate the presence of a strong, persistent thermally driven circulation within the boundary layer, with the cloud circulation being secondary. Collapsing turrets appear to be responsible for the significant detrainment that occurs at mid levels within one simulated cloud. Transport by downdrafts is significant throughout the cloud and subcloud layers.

The boundary layer and cloud circulations dry the subcloud layer, with significant detrainment of moisture occurring in the upper portion of the boundary layer. Strong apparent moistening in the upper half of one cloud is driven by mean vertical transport of moisture toward the detrainment layer aloft, although detrainment of moisture at midlevels is a comparatively small component of the apaprent moistening. Storage of moisture is found to be an important effect.

This paper is the first in a three-part series in which a three-dimensional numerical model is run at high resolution to simulate cumulus congestus clouds in three dimensions with the principal goal of understanding the mechanisms associated with entrainment and detrainment. The clouds are contained within a nested grid having a 50-m uniform grid spacing; the model does not allow precipitation or ice formation and achieves saturation through bulk condensation. The prescribed environment is that associated with nonprecipitating New Mexican cumulus clouds observed on 9 and 10 August 1987.

The convection is initiated using continuous surface heating, including a central Gaussian component to represent the effects of an isolated mountain range. Several regions of concentrated surface heating are used during the first hour to condition the environment. The turbulent motion thereby introduced into the boundary layer is crucial for the accurate simulation of the clouds.

The simulated clouds extend vertically up to 4 km, and model results generally agree with aircraft observations in quantities such as cloud base and top height and the presence or absence of pronounced detrainment layers at midelevels. Further, the pulsating nature of the convection, in which the clouds strengthen and decay over periods of several minutes, is also similar to observations. The cloud-top height is generally not correlated with the level of neutral buoyancy for hypothetical parcels ascending undilute.

Spatial resolution at least as fine as that used here appears necessary in order to capture the details of cumulus entrainment, although clouds simulated on a single coarse grid exhibited a substantial degree of similarity to their nested grid counterparts and were at times somewhat more vigorous.

Flights through the central regions of thunderstorms were made over New Mexico on 6 and 15 August 1977 with the ONR/NMIMT Schweitzer aeroplane which carried equipment designed to measure all three components of the electric field, and the charge, $Q$, and diameter, $d$, of individual precipitation elements. On the earlier day, information was also obtained with: a rain-gauge network surrounding Langmuir Laboratory; a 3 cm radar; an acoustic system for locating lightning channels; a ground-based field-change meter.

The first cell on 6 August produced precipitation at the ground but no lightning. Vertical fields, $E_z$, of up to about 50 kV/m and precipitation charge densities $\rho_p$ of up to $-0.5 $C/km³ were recorded within the cloud. The second cell, which grew as the first one decayed, produced 7 lightning strokes in 9 minutes during which time the radar revealed vigorous vertical growth in a narrow zone containing precipitation.

Thunder reconstructions showed the acoustic sources for the first flash of this cell to be very near the top of the cloud at an altitude of 10 km a.s.l. The subsequent flashes produced acoustic signals from progressively lower in the cloud. When the radar echo reached its maximum height lightning activity ceased. $E_z$ values of up to about 50 kV/m and $\rho_p$ values down to -1 C/km³ were measured.

$\rho_p$ was consistently negative, individual charges being less than $\pm 40$pC. $Q$ values were within the inductive limit for a thundercloud at breakdown but no systematic relation between $Q$ and $d$ was found.

Six penetrations were made through the thundercloud of 15 August, which produced only two lightning strokes. The $E_z$ records were indicative of a $(\pm)$ dipole located near the cloud top, at around $-13\C$. Fields of up to about 100 kV/m and $\rho_p$ values (positive and negative) of around 5 C/km³ were measured. $Q$ values of up to $\pm 250 $pC were recorded, with charges around $\pm 50 $pC being commonly found. No systematic $Q-d$ relation was revealed, and smaller precipitation particles frequently carried charges (positive or negative) in excess of the inductive limit.

On both days estimated precipitation rates were of order 10 mm/h and on most occasions the pilot reported precipitation particles to be either 'ice' or 'mixed liquid water and ice'.

Measurements of charge versus size of drops within warm thunderstorm clouds indicate that the mean specific charge per gram decreases dramatically as the size increases. Because coalescence between smaller and larger drops occurs with an approximately geometric cross section, the charge neutralization length (the distance in which the smaller drops of one charge sign could neutralize the charge of the larger drops by coalescence) becomes small in comparison with cloud dimensions and hence ensures that any effective charge separation mechanism rapidly approaches steady state. The downward, steady-state, large drop mass flux averaged over the cloud is usually less than the upward, convected, small drop mass flux. At the interior of the cloud, quasi steady-state charge separation requires an equal and opposite charge flux carried by smaller and larger drops. The two conditions are inconsistent with current drop charge versus size measurements by several orders of magnitude. Currently, the assumption of equal and opposite charge flux is not substantiated and hence charge separation by differential 'falling' of larger and smaller drops within the cloud should be a small contribution to the total charge transport.

Expressions are derived which relate the terminal fall speed of hydrometeors to their size and density and to the environmental parameters of the air, and for extrapolating terminal velocity measurements under known conditions to other conditions and altitudes. The formulation is an extension of previous work and is based on the assumption that on average the drag coefficient is related to the Reynolds number by a simple power law over a reasonable range of Reynolds numbers. This is shown to imply a power law dependence between the terminal fall speed and diameter of a particle, as is often observed, and also to be consistent with empirical power law relations between the Best and Reynolds numbers of falling particles. A set of terminal fall speed parameters is presented which is representative of a variety of precipitation types.

Aircraft, radar, and surface observations were used to study the relationship between precipitation development and the onset of electrification in thunderstorms which formed near or over the Magdalena Mountains of New Mexico. The study included storms which were electrically active as well as ones in which no electrical enhancement was observed. Electric fields inside these clouds showed negligible enhancement and did not exceed 1 kV/m until reflectivities at 6 km above mean sea level (msl) (about $-10\deg$C) exceeded approximately 40 dBZ and cloud tops exceeded 8 km. The onset of electrification during or immediately after convective growth within the cloud.

Airborne electric field measurements in two small thunderstorms in New Mexico show the existence of a narrow region of charge in each storm during the early stage of electrification. In one case a net negative region of charge was observed at about 7 km ($-12\deg$C) about 1.5 km below the radar cloud top, when the electric field was only 600 V/m, and could be accounted for by total charge of $-0.01$C with a maximum net space charge density of $-0.15$nC/m³, assuming spherical symmetry. This region of charge was about 500 m across and appears to have been associated with an updraught-downdraught transition zone. In the other cloud, a region of net positive charge, also about 500 m across, was detected at 7.7 km ($-20\deg$C) about 500 m below the radar cloud top, when the electric field was about 2000 V/m. In both regions of charge, supercooled liquid water and ice particles including graupel were present, and ice particle concentrations, sizes, and collision rates were at a relative maximum, suggesting that the charge generation occurred via a precipitation-based mechanism.

During the Summer of 1976 the ONR/NMIMT research aeroplane was employed in studies of the electrical properties of thunderstorms. Flight through clouds were made in Florida, as part of the Thunderstorm Research International Project, and in New Mexico. The most important measurements were of electric field and the charge, $Q,$ and size $d,$ of individual precipitation elements. A novel device was constructed for the $Q$ and $d$ measurements. The charge carried on a particle passing through a metal cylinder was sensed by induction, and its size by a shadowgraph technique involving a linear array of photo-diodes. The penetrations were generally through the lower regions of the clouds.

The major findings of the studies in New Mexico were as follows:

1. Volume charge densities on precipitation, $\rho_p$, were often around $-5$ nC $m^-3 over horizontal distances of several kilometres. $\rho_p$ was almost always negative, but positive charge densities, of lower magnitude, were occasionally observed over shorter distances. The major contribution to the measured values of $\rho_p$ was made by particles of size around 1 mm, or smaller.

2. Simultaneous measurements of $Q$ and $d$ showed that no simple relationship existed between them. Charges of about 100 pC were commonly observed on particles around 1 mm in size. These are much too high to be explicable in terms of the inductive theory.

3. Positive and negative charges were found to coexist, except when the precipitation rate, $p$, was very low. However, charge of one sign (almost invariably negative) was always strongly dominant.

4. Values of $p$ could be estimated crudely from the $d$ pulses. In regions of high $\rho_p$ they were rarely in excess of 10 mm/h; on some occasions when $\rho_p$ was substantial $p$ was below 1 mm/h.

Our primary conclusions are that in the clouds studied substantial currents were often carried on precipitation, and that the charges on individual precipitation elements are not explicable in terms of the inductive mechanism of thunderstorm electrification.

Typically, 50-70% of the total annual precipitation in New Mexico can be produced by convective thunderstorms during the period June through September. These thunderstorms are accompanied by intense lightning and characteristically produce heavy, localized rainfall resulting in high spatial variation in precipitation inputs. During other months precipitation over the entire Sevilleta (10(5) ha) often occurs from broad-scare storm systems and is much less spatially variable on a per-storm basis. Summer precipitation is a primary factor driving plant productivity as well as influencing nutrient cycling, herbivore activity, and detritivore activity. Knowledge of the timing, location, and amounts of precipitation is important in planning or monitoring research activities and spatial modeling of the dynamics in this semiarid region. Technology exists for locating cloud-to-ground lightning strikes that has the potential to locate these intense precipitation events, quantify the volume of water associated with them, and document the spatial and temporal variability of this phenomenon over large areas. Near real-time analysis capability can identify areas receiving precipitation that will experience rapid vegetation growth in this semiarid region. This study developed algorithms relating lightning and precipitation quantity and used lightning location to determine rainfall depth and distribution for areas in New Mexico. There was a significant correlation between rain-gauge measured precipitation and lightning within a 3-km radius of the gauge location, with best predictions occurring from regressions that included lightning strikes and relative humidity. Average precipitation volume per cloud-to-ground lightning strike averaged 36190 m(3) for the 3 km radius circle, resulting in an average rainfall depth of 1.3 mm per lightning strike. Lightning location technology, combined with a Geographic Information System (GIS), defined the spatial and temporal resolution of these intense, summer precipitation patterns and provided a more detailed estimate of total precipitation and precipitation distribution than was provided by the sparse network of precipitation gauges. Combining this information with satellite sensing of vegetation growth (e.g., greenness index) can identify causal mechanisms for temporal and spatial patterns in short-term vegetation processes (e.g., primary production) and long-term vegetation dynamics for this area.

A variety of new measurement techniques and coordinated observations are beginning to reveal the electrical nature of thunderstorms.

It is possible under some circumstances that the accuracy of data on electrical parameters inside thunderclouds obtained by use of instrumented balloons may sometimes be adversely affected by the balloon platform itself. If the balloon rigging becomes conductive, polarization of the balloon system, corona emissions, and electrical generation of charged water particles may act to alter the parameters that are being measured. Laboratory measurements of conduction currents flowing along nylon monofilaments, sometimes used as insulating supports for instruments carried beneath balloons, revealed at large conductance when the filaments were wet or exposed to high humidity and the ambient temperature was above 0-degrees-C. The conductance of a piece of monofilament at high humidity increased by more than 4 orders of magnitude as the air temperature rose from -15-degrees to 20-degrees-C. The conductance also depended strongly on the monofilament's surface cleanliness. At 20-degrees-C in high humidity, currents of up to 0.3-mu-A flowed along a piece of monofilament in response to a potential gradient of 2000 V/m. Application of large potential gradients to a monofilament continually wetted, as might be expected as a worst-case scenario in clouds or rain, produced corona emissions and electrical generation of charged drops. These results indicate that especially in the warmer, lower regions of clouds, currents may flow in the rigging of balloon-borne instrumentation sufficient to reduce the ambient atmospheric electric fields and to generate a population of artificially charged water particles. In order to ensure credibility of electrical measurements in such situations, it is essential that conduction along the rigging be minimized by the use of hydrophobic materials, and it is desirable that means of monitoring leakage currents be included in the instrumentation.

The temporal variation of electric field components deduced from measurements made with airplanes penetrating electrified clouds is often complex, especially when the airplane experienced strong electrical charging. However, unusually simple electric field variations were obtained for penetrations involving severe charging of an airplane on flights over Kennedy Space Center, Florida, on August 19, 1989. Analysis of these results suggests that plumes of ions were emitted from the airplane. The electric field from these plumes was more intense than the ambient field from the cloud at the location of the aft electric field sensor. As a consequence the deduced component of the ambient electric field in the direction of flight was severely distorted. These findings emphasize (1) the importance of careful evaluation of electric field data obtained with airplanes and (2) the need for improved measurements.

Three airplanes --- a sailplane, a piston powered sailplane, and a twin turboprop --- have been instrumented to measure electric fields inside electrified clouds and the net electric charge on the airplanes. In unelectrified clouds the powered airplanes become negatively charged during collisions with liquid cloud water droplets whereas the sailplane does not. In thunderstorm clouds, several airplane charging mechanisms are found to operate. These involve collisions with liquid water droplets of the cloud and the shedding of polarization charge in the presence of strong electric fields. For these charging processes, the sign of the acquired charge depends on the sign of the component of the atmospheric electric field along the direction of flight. When the amounts of charge on the powered airplanes are small, the engine exhaust acts to discharge the airplane, while for larger charges, corona emission becomes the predominant mechanism of discharge.

A simple instrument that detects ice particles has been developed for use in airplane studies of thunderstorms. Although sophisticated instruments are available for imaging atmospheric ice particles, the spatial resolution of the particle concentration determined from their data is limited by the small size of the sample area. The ice detector described here provides a real-time indication of the presence of ice crystals within and in the vicinity of clouds, and provides an approximate measure of their concentrations with excellent spatial resolution. This device, which is simple and inexpensive, has been flown for five summers in New Mexico thunderstorms.

Lightning discharges were investigated with high time-resolution equipment on both electric-field and electric-field-change meters.

The analysis of the electrical records reveals that the late stages of intracloud discharges are very similar to those of cloud-to-ground discharges during the periods between sucessive return strokes (junction process) and during the period after the last return stroke (final process). In contrast, the initial portion of the field change of an intracloud discharge bears little or no resemblance to the initial portion of the leader field change of a discharge to ground. It is suggested that the difference in the initial breakdown characteristics results from variations in the relative populations of water drops and ice particles as they affect the internal impedence of the region of cloud where breakdown occurs.

The difference in the initial field-change characteristics of intracloud and cloud-to-ground discharges is so distinct that from the first 10 ms of the electric-field-change record one can predict with over 95% certainty whether a discharge will reach ground or remain within the cloud.

Observations of thunderstorms with a dual channel circular-polarization radar have provided dramatic indications of the buildup of the electric field inside the storms and of the sudden collapse of the field at the time of lightning. The indications are obtained by coherently correlating the simultaneous returns in the right- and left-hand circular polarization channels of the radar, and follow up on the pioneering observations of this type by Hendry and McCormick (1976). The correlation is estimated and displayed in real time and the results enable one to predict when a storm has the potential for producing a lightning discharge, and often to anticipate the occurrence of individual discharges. The observations detect the presence of electrically aligned particles, believed to be small ice crystals, which are aligned by the electrostatic field of the storm. The aligned particles cause the radar signal to become progressively depolarized as it propagates through an alignment region, giving rise to correlated right- and left-circular polarization echoes. The alignment direction can be determined from the phase of the correlation and is found to be predominantly vertical, indicating a similar electric field orientation. Weaker horizontal alignment is often observed immediately following lightning discharges, consistent with the idea that the aligned particles are ice platelets which fall with horizontal orientation due to aerodynamic forces. The observations have been found to reveal the onset of strong electrification in developing storms and to indicate when decaying storms no longer have the potential to produce lightning. By compensating for signal-to-noise effects, the variation of the depolarization with range can be determined. This provides detailed pictures of the alignment regions which could be used as tracers of ice crystal populations in storms. The pictures also show the spatial variation of the alignment directions, raising the possibility of remotely mapping the storm electric field structure. Finally, the depolarization rate results readily enable one to distinguish between liquid and solid precipitation in the storms.

Sources of charge for the individual strokes of four multiple-stroke flashes to ground have been determined, using measurements of the electrostatic field change obtained at eight locations on the ground beneath the storm. The resulting charge locations have been compared to 3-cm radar measurements of precipitation structure in the storm. The field changes of individual strokes were found to be reasonably consistent with the lowering to ground of a localized or spherically symmetric charge in the cloud. The centers of charge for successive strokes of each flash developed over large horizontal distances within the cloud, up to 8 km, at more or less constant elevation between the $-9\deg$ and $-17\C$ environmental (clear air) temperature levels. Comparison with the radar measurements has shown that the discharges developed through the full horizontal extent of the precipitating region of the storm and appeared to be bounded within this extent. In one instance where cellular structure of the storm was apparent, the strokes selectively discharged regions where the precipitation echo was the strongest. Vertical extent of the stroke charge locations was small in comparison with the vertical extent of the storm. The field changes in the intervals between strokes have been found to exhibit many of the features which Malan and Schonland used to infer that ground flashes discharge a nearly vertical column of charge in the cloud. This and other evidence is used to show that their observations, which were made at a single station, could instead have been of horizontally developing discharges. The interstroke field changes have been analyzed using a point dipole model and found to correspond to predominantly horizontal charge motion that was closely associated with the ground stroke sources for the flashes. The interstroke activity served effectively to transport negative charge in the direction of earlier stroke volumes and often persisted in the vicinity of an earlier stroke volume, while subsequent strokes discharged more distant regions of the cloud. Long-duration field changes that sometimes preceded the first stroke of a flash have been analyzed and found to correspond to a series of vertical and horizontal breakdown events within the cloud, prior to development of a leader to ground. These events were associated in part with the negative charge region that became the source of the first stroke and effectively transported negative charge away from the first stroke charge volume and from the charge volumes of subsequent strokes. Several continuing current discharges were found also to progress horizontally within the cloud and sustained currents in the range of 580 A to less than 50 A. The continuing current field changes were consistently better fitted by the monopole charge model than the field changes of discrete strokes within the same flash.

On August 15, 1984, we made balloon measurements of the electric field and of the charge and vertical velocity of precipitation particles in a thunderstorms over Langmuir Laboratory. Substantial quantities of positively charged precipitation, comprising a lower positive charge are particles found higher in the cloud seem to have generated electrical energy in the way suggested in the typical theory of thunderstorm electrification.

The total charge density and the precipitation charge density inferred from six balloon soundings of electric field and precipitation charge are compared in order to examine the distribution of cloud charge density in small New Mexican mountain thunderstorms. Cloud charge density magnitudes were typically less than 2.0 nC m^-3, and the maximum cloud charge densities were -2.1 and +6.7 nC m^-3. The cloud charge density made a significant contribution to the total charge density in most of the thunderstorm depth, especially in the upper positive and upper negative charge regions, where all the charge was typically on cloud particles. The largest positive cloud charge densities were found just above the main negative charge region. In the lower part of the main negative charge region, the cloud charge density was the major contributor to the total charge density, and the precipitation charge played a minor role. The cloud charge density was often zero in lower positive charge regions but was 20-50% of the total charge density in two cases. Between the upper positive charge region and the bottom of the main negative charge region the cloud charge density changes polarity and tends to become increasingly negative with decreasing altitude.

Instruments that measure the intense electric field strengths in thunderclouds (-100 kV/m) are designed to minimize the production of ions by small electrical discharges (coronas) emanating from the instruments themselves. The nearby charge of these ions would unpredictably disturb the natural field of the cloud. In an attempt to assess this disturbance, two different instruments (one carried by a rocket and one carried by a balloon) were launched on two occasions into thunderstorms. In spite of differing trajectories, the soundings were similar, which gives us some confidence in both instruments. In addition, the measurements revealed some interesting features of the two storms. Each storm appeared to have six significant and distinct regions of charge. The balloon soundings also revealed that lightning flashes temporarily increased the electric field strength above the thunderclouds (at altitudes from 9.7 to 14.3 km) by amounts up to 10 kV/m, after which the fields decayed away in 50 to 125 s. One pair of ascent and descent rocket soundings, separated in time by a maximum of 60 s and horizontally by 1 to 3 km, showed little change in the thunderstorm electric field between ground and 7.5 km altitude.

Substantial quantities of negatively charged precipitation were found in the main negative chage center of a small New Mexico thunderstorm. The charge density of the negative precipitation was equal to or greater than the deduced total charge density of the region.

We made one balloon sounding of the electric field in each of two dying thunderstorms. Both balloons were launched 20 min after the storm's last lightning flash. Each sounding revealed substantial residual electrification. The peak magnitudes of electric field in the two dying storms were 35 and 71 kV/m. The charge structure in both storms appeared to consist of one internal negative charge region with positive screening layers at the upper and lower cloud boundaries. One or two negative charge regions were found below cloud base. The main charge regions had inferred densities of about 1 nC/m³ and were hundreds of meters thick.

Two electric field soundings through thunderstorm anvil clouds show similar charge structures: negatively charged screening layers on the top and bottom surfaces, a layer of positive charge in the interior, and one or two layers of zero charge. Both anvil clouds were strongly electrified: the peak magnitudes of the electric field in the two storms were 70 and 90 kV/m, respectively. The nonzero layers had charge densities comparable to those found in the cores of thunderstorms, ranging in magnitude from 0.4 to 2.7 nC/m³. Layers varied in thickness from 300 to 2000 m. The positive charge probably originated in the main positive charge region normally found at high altitudes in the core of thunderclouds. Transporting positive charge from the storm core to the anvil may influence the ratio of intracloud to cloud-to-ground lightning flashes and the rate of generation of charge in the core. The negatively charged layers probably were screening layers, resulting from the discontinuity in the electrical conductivity at the cloud boundaries. The lower negative screening layer appeared to be carried toward the storm core by winds below and at the lower anvil boundary.

We designed an instrument to measure the charge and vertical velocity of individual precipitation particles inside thunderclouds. A balloon carried the particle charge instrument, an electric field meter, and a standard meteorological radiosonde upward into thunderclouds over Langmuir Laboratory in central New Mexico. During one balloon flight the instruments encountered two regions of positive charge below the main negative charge center. We identify these positive regions with the lower positive charge centers that have been desctibed in the literature for many years. We find the following points: (1) One region had an estimated total charge of 0.4C. The other had 2 C. (2) The charge resided on precipitation particles. The particles' charges typically ranged between 10 and 200 pC, but a few particles had charges up to 400 pC. Their diameters lay between an estimated 1--3 mm. The charges were too large to be explained by the polarization induction mechanism. We favor the hypothesis that lightning provided the positive charge in the lower positive charge centers. (3) The motion of the lower positive charge centers enhanced the electrical energy of the storm, but their contribution to the overall electrical budget was small. (4) The field excursions (at the ground) associated with precipitation (FEAWPs) described by C.B. Moore and B. Vonnegut are probably caused by lower positive charge centers descending on precipitation. The larger (2 C) lower positive charge center caused a FEAWP. Negatively charged precipitation particles passed through our instrument near the top of its trajectory just before the balloon was struck by lightning. The charge density on precipitation particles was substantial, but we do not have enough information to comment on the role the particles may have had in generating the main region of negative charge.

Direct emissions of nonmethane hydrocarbons, monocarboxylic acids, and low molecular weight carbonyl compounds were measured from vegetation typical to central New Mexico. These species included quaking aspen, cottonwood, Gambel oak, Douglas fir, Engelmann spruce, Rocky Mountain juniper, pinyon pine, and ponderosa pine. The hydrocarbon emissions from most of the coniferous trees were dominated by alpha-pinene. In general, alpha-pinene emissions were 100-10 000 ng g(-1) h(-1) and displayed the expected temperature dependence. Other identified hydrocarbons included isoprene, camphene, beta-pinene, myrcene, Delta(3)-carene, and d-limonene, The deciduous trees as well as the spruce and fir trees showed isoprene emission rates of 100-100 000 ng g(-1) h(-1). Formaldehyde and acetaldehyde were the most common low molecular weight carbonyl compounds measured. The carbonyl emissions averaged 50-1660 ng g(-1) h(-1), depending an the compound and the trees species. Unlike the hydrocarbons, the carbonyl emissions displayed little correlation with enclosure temperature. Formic acid emissions averaged 15-920 ng g(-1) h(-1), and acetic acid emissions averaged 50-1300 ng g(-1) h(-1). As with the carbonyls, poor correlation was found between the acid emissions and the enclosure temperature. The deciduous trees were found to have average (mass-based) emissions of 98% hydrocarbons, 1% carbonyls, and 1% organic acids. The coniferous trees averaged 80%, 8%, and 12%, respectively.

This research consists of a laboratory study and a field study. The laboratory research reports the formation of NOx from a point to plane corona discharge. Discharge polarity and relative humidity determined the amount of NOx that was produced. The positive point discharge caused more NOx to form than the negative point discharge. For both polarities NOx production showed a nonlinear increase with current. Relative humidity enhanced the NOx formation for both polarities; Ln each case, the amount of NOx formed was comparable to the quantity of N2O produced from corona discharge. The research also reports the results from a field study that measured the amounts of O-3 and NO2 produced by corona discharge during a thunderstorm. The study found that the ambient concentrations of O-3 and NO2 increased several fold due to corona discharge and returned to original levels after the thunderstorm. Copyright (C) 1996 Elsevier Science Ltd

The nature of visible, horizontally stratified lightning channels propagating over large distances near the cloud base during the decaying stage of a storm (also called ''spider'' lightning) was investigated. The study was effectuated through the use of the coordinated observations of a VHF interferometer, a high-speed image-intensified video system, measurements of electric and magnetic fields, and optical transients. Spider-lightning events were found to be negative leaders similar to stepped leaders in negative cloud-to-ground flashes, with a similar average speed of propagation horizontally of 2-4 x 10(5) m s(-1) Being slow negative leaders, spider-lightning events are part of intracloud flashes and positive cloud-to-ground flashes occurring prior to and during the inverted (fair weather polarity) phase of the End of the Storm Oscillation in the ground electric field. Spider lightning is characterized by both the pulsing luminosity at the tips of its branched channels and the continuous luminosity (for tens to hundreds of milliseconds) which is maintained by the continuing current flow. The interferometer produced mapping of radiation sources closely resembling the spider-lightning channels (negative leaders) but only;a weak trace of radiation sources associated with positive leaders to ground. Both the video images and a few radiation sources of positive leaders were obtained within 1 ms of the leader's ground attachment. The interferometer, however, failed to map fast negative leaders that occurred intermittently during the spider-lightning events.

A six-stroke cloud-to-ground lightning flash has been studied using observations from a high-speed video camera (1000 frames/s) and a VHF radio interferometer (1-$\mu$s time resolution), as well as additional electric, magnetic, and optical measurements. The flash produced strokes along two channels to ground and a long (550 ms) continuing current. The video observations provided time-resolved pictures of stepped and dart leaders, short and long continuing currents, and M components during the continuing currents, and complemented and confirmed the interferometer observations of flash structure and development. The M components were initiated by fast negative streamers inside the cloud which propagated into the conducting channel of the continuing current and subsequently brightened the channel to ground. We call this sequence an M event. Dart leaders and the fast streamers of M and K events were found to be significantly brighter than stepped leaders and continuing currents to ground. A number of streamers did not initiate M events that were identical to in-cloud K streamers. Analysis of the M and K event occurrences indicated that the conducting channels of the continuing current both expanded and contracted with time. An M-type event was also observed during a dart leader. It is proposed that the channel multiplicity within the flash resulted from cutoff of the channel to ground while charge continued to flow down the channel from the stroke source region, stranding negative charge along the channel. Dart-stepped leaders (such as occurred during the third stroke) are similarly explained. Because the stranded charge is observed to be greatest for initial strokes, new channels to ground of stepped and dart-stepped leaders are expected to follow initial strokes, as is usually observed. The video and electric field observations indicate that all return strokes have at least a short continuing current, of the order of one or a few milliseconds. The results also reinforce the well-known observation that long continuing currents tend to follow relatively weak return strokes.

The South Dakota School of Mines and Technology armored T-28 airplane has been used to collect electric field data inside thunderstorms since 1986. However, the derivation of the E-x field component in the flight direction has been problematic, and the accuracy of other field components derived has been suspect when heavy precipitation was encountered. The problems are due to (1) poor understanding of when and where corona ions are emitted from the airplane and how they affect the measurements, (2) lack of a complete and advanced calibration of the electric field meters' response to the ambient electric field components and airplane charge, and (3) poor understanding of this particular airplane's charging characteristics in the heavy precipitation and strong electric field environment of severe thunderstorms. By using data obtained during intercomparison/formation flights of the T-28 with another well-calibrated airplane in 1997, a complete calibration matrix for the T-28 is presented in this paper. The calibration matrix provides improved and more robust estimates of all electric field components E-x, E-y and E-z. The data from the intercomparison flights are used to characterize the corona emission phenomena and its effects on the meter measurements. A methodology to detect emissions and their effect on the measurements is presented. Findings from this work and work from the Special Test Vehicle for Atmospheric Research (SPTVAR) at New Mexico Institute of Mining and Technology indicate that electric meters should not be located downstream of any airplane's propeller, the most likely source of corona ion emissions in large ambient electric fields. Alternative computations for field components and techniques for calculating airplane charge are also presented. In addition, we show that when flying through heavy precipitation, the effect of precipitation charging on a field meter's reading is small in comparison to the effect of typical ambient electric fields found in thunderstorms.

The growth of electrical energy in a thundercloud prior to a lightning flash can be studied by measuring the electric field. The use of airplanes to measure the electric field has been a problem because the intense electric field of a thunderstorm causes electric breakdown in air at sharp metallic paints and edges. The electrical charge produced by the breakdown can be confused with the charge in the thunderstorm itself, causing large errors in the measurement. Electric field sensors on the fuselage of a single-engine airplane can give erroneous measurements when the electric field amplitude is as low as 10 kV/m. However, new electric field sensors in peas, under the wings away from the fuselage appear to give correct measurements up to about 150 kV/m, the most intense field we have encountered. A flight around a rainshaft beneath a thunderstorm shows how the electric field, deduced from the new pod meters under the wings, differs from the field deduced from meters on the fuselage.

With the use of a NaI scintillation detector, bursts of radiation with energies in excess of 1 MeV were recorded at a mountain-top observatory immediately before three, nearby cloud-to-ground, negative lightning strikes. Coincident recordings of the electric field changes due to the discharges showed that, in each case, the bursts began between 1 and 2 milliseconds before and continued until the onset of the first return stroke. This radiation was associated with approaching stepped-leaders and may have influenced their development.

Following Benjamin Franklin's invention of the lightning rod, based on his discovery that electrified objects could be discharged by approaching them with a metal needle in hand, conventional lightning rods in the U.S. have had sharp tips. In recent years, the role of the sharp tip in causing a lightning rod to act as a strike receptor has been questioned leading to experiments in which pairs of various sharp-tipped and blunt rods have been exposed beneath thunderclouds to determine the better strike receptor. After seven years of tests, none of the sharp Franklin rods or of the so-called ''early streamer emitters'' has been struck, but 12 blunt rods with tip diameters ranging from 12.7 mm to 25.4 mm have taken strikes. Our field experiments and our analyses indicate that the strike-reception probabilities of Franklin's rods are greatly increased when their tips are made moderately blunt.

Although lightning rods have long been used to limit damage from lightning, there are currently no American standards for the shape and form of these devices. Following tradition, however, sharp-tipped Franklin rods are widely installed despite evidence that, on occasion, lightning strikes objects in their vicinity. In recent tests of various tip configurations to determine which were preferentially struck by lightning, several hemispherically tipped, blunt rods were struck but none of the nearby, sharper rods were ''hit'' by lightning. Measurements of the currents from the tips of lightning rods exposed to strong electric fields under negatively charged cloud bases show that the emissions consist of periodic ion charge bursts that act to reduce the strength of the local fields. After a burst of charge, no further. emissions occur until that charge has moved away from the tip. Laboratory measurements of the emissions fi om a wide range of electrodes exposed to strong, normal-polarity thunderstorm electric fields show that positive ions are formed and move more readily over sharp-tipped electrodes than over blunter ones. From these findings, it appeals that the electric held rates of intensification over sharp rods must be much greater than those over similarly exposed blunt rods for the initiation of upward-going leaders. Calculations of the relative strengths of the electric fields above similarly exposed sharp and blunt rods show that although the fields, prior to any emissions, are much stronger at the tip of a sharp rod, they decrease more rapidly with distance. As a result, at a few centimeters above the tip of a 20-mm-diameter blunt rod, the strength of the held is greater than that over an otherwise similar, sharper I od at the same height. Since the field strength at the tip of a sharpened rod tends to be limited by the easy formation of ions in the surrounding ail; the field strengths over blunt rods can be much stronger than those at distances greater than 1 cm over sharper ones. The results of this study suggest that moderately blunt metal rods (with tip height-to-tip radius of curvature ratios of about 680:1) are better lightning strike receptors than are sharper rods or very blunt.

Tests of influence mechanism explanations for thundercloud electrification have been attempted by modifying the initial atmospheric electrical conditions beneath some convective clouds. The modification was accomplished by the release of negative space charges into the air from elevated wires connected to a high-voltage power supply. Anomalous distributions of charge were observed in some clouds with low bases that grew over the charge source. The results of the experiment suggest that influence mechanisms may be operative in the electrification of our mountain clouds and that, normally, the natural, positive space charges in the sub cloud air may initiate the electrification process.

Franklin invented lightning rods with the hope that they would dissipate thunderstorm electricity and thus prevent lightning from striking. His invention was based on his findings that sharpened metal needles would allow electricity to flow silently through the air, away from highly charged objects. When his rods were used, however, instead of preventing lightning, they were sometimes "struck" and became part of a lightning path to earth.

An analysis of the physics involved suggests that:

(a) The flow of electricity from sharpened conductors at the earth's surface does not dissipate thunderstorm electricity sufficiently to prevent lightning.

(b) The ionization and point discharges around the tip of a sharpened lightning rod limit the strength of the local electric field and reduce the probability of a lightning strike to the rod. The sharpened rod thus acts to protect itself against lightning discharges, but its protection does not extend to other objects in its vicinity. While a sharpened rod does not provide a preferred lightning path to earth, it can be used if no better paths are avilable.

(c) Elevated, blunt rods or horizontal conductors, suitably connected to earth, can provide better lightning paths to earth and therefore, better protection to structures in their vicinity than do sharpened rods.

(d) The connections from elevated conductors to earth need to be the most direct possible, with no abrupt changes in direction; impedance discontinuities created in down conductors at sharp bends cause reflections of lightning transients and may produce side flashes to other objects in their vicinity.

Neglect of the angle dependence of rebound probabilities for cloud droplets, colliding with hail pellets has led to an overestimate of the amounts of charge that can be separated with the Elster-Geitel mechanism. Application of the wind tunnel results obtained by Aufdermaur and Johnson suggests that the actual amounts of charge transferred may be less than 20% that assumed by ignoring the angle dependence. Observations of the charges carried by hail and rain particles within the bases of thunderclouds support these lower estimates and raise questions about the adequacy of the precipitation processes in explanations of thundercloud electrification.

Observations of thunderstorms in New Mexico were made with a vertically-scanning, 3-cm radar on a mountain top. Prior to a cloud-to-ground lightning discharge nearby, the radar echo overhead was usually quite weak, indicating low intensities of precipitation there. Following the lightning it was observed sometimes that in the region of the cloud where the discharge occurred the radar echo intensity rapidly increased, and shortly thereafter a gush of rain or hail fell nearby.

These studies confirm earlier radar observations, made by the authors at Grand Bahama Island, B.W.I., in whch it was found that lightning is often followed in the cloud by a rapidly intensifying echo and then by a gush of rain at the ground. The increases in radar reflectivity in small volumes of the cloud following lightning suggest that the electric discharge is influencing the size of particles in the cloud.

An analysis indicates that within 30 seconds after a lightning discharge, the mass of some droplets may increase as much as 100-fold as the result of an electrostatic precipitation effect.

Observations of summer thunderstorms developing over Mt. Withington, New Mexico, show that electrification begins early in the cloud's development. Measurements within and above the cloud show that electric charge accumulations similar in polarity to those of the mature storm begin to form before any echo can be seen with an X-band radar and before any electrical perturbations can be detected on the mountain summit beneath the cloud.

Measurements of potential gradient were made within the cloud with radiosondes supported on tethered balloons. It was found that the gradient within the cloud was far larger than that outside the cloud and that it reached values as high as 20 V/cm before the appearance of the radar echo. Measurements made from an airplane flying over the top of the growing cloud showed that here the fair-weather potential gradient reversed before any radar echo could be seen. The rapid rise of the cloud tops and the tension on the line holding the tethered balloon showed that the development of the initial electrical activity was closely related to convective activity.

The initial radar precipitation echo within the cloud was frequently in the form of a hollow, inverted cup that usually filled in and became completely reflecting in a few minutes.

The negative charge involved in lightning flashes to ground is found to be distributed in a manner strongly dependent upon the direction of movement of the storm, and does not, in general, constitute a nearly vertical column as proposed by Malan and Schonland. Based on a study of electric field-changes measured at two stations 10 km apart involving 539 return strokes from 84 flashes in 10 storms, we conclude that the horizontal component of the in-cloud channel on the average exceeds the vertical component, and points in the dirction of storm motion.

An analysis is given of the five methods of Malan and Schonland, on the basis of which we suggest that significant horizontal components are also compatible with their observations of a vertical column, and that the 'nearly vertical' aspect of the charge distribution has been over-emplahsized.

A new 3D model with explicit liquid- and ice-phase microphysics and a detailed treatment of ice nucleation and multiplication processes is applied to study ice formation and evolution in cumulus clouds. Simulation results are compared with in situ observations collected by the National Center for Atmospheric Research King Air aircraft in a cloud over the Magdalena Mountains in New Mexico on 9 August 1987. The model reproduces well the observed cloud in terms of cloud geometry, liquid water content, and concentrations of cloud drops and ice particles (IP). Primary ice nucleation is shown to produce IP in concentrations on the order of 10³m^-3 (1 L^-1) once the cloud top reaches $-10 /deg$ to $-12 /deg$C. At mature and early dissipating stages of cloud development, ice production is dominated by the rime-splingering (Hallett-Mossop) mechanism, which in some regions generates up to 5 x 10⁴ m^-3 (50 L ^-1} IP in about 10 min. The predicted maximum of IP concentration is in agreement with observations. The sampling techniques used in the field study, however, do not provide an adequate estimate for the splinter production rate, which exceeds 100 m^-3 s^-1 in the model.

The Met Office Cloud Resolving Model (CRM) and the UMIST Explicit Microphysics Model (EMM) have been employed in the analysis of data from airborne studies of a multi-thermal cumulus cloud which developed over New Mexico in the summer of 1987. The principal goal was to establish a quantitative understanding of the observed development of glaciation of this cloud.

The EMM was utilized in a series of tests designed to assess the sensitivity of cloud glaciation via the Hallett-Mossop (H-M) process to cloud parameters such as the concentration of cloud condensation nuclei, the cloud-base temperature, entrainment, and the freezing and splintering of upercooled raindrops. These tests with the EMM demonstrate that reductions in the mean droplet diameter can inhibit the rates of H-M splinter production and auto-conversion, reducing the rate of accumulation of precipitation at the ground and reducing the concentration of ice particles. The warm-rain process in the EMM is fundamental to the production of graupel, H-M splinters and precipitation.

Good agreement was found between the predictions of the CRM and the available dynamical and microphysical field observations. Analysis of results from both models indicated that the cloud glaciation is explicable in terms of the H-M process, with ice production being dominated by the freezing of supercooled raindrops in the H-M band, and the immediate and continuous production of ice splinters as supercooled droplets freeze onto them.

Measurements of the levels of Be-7 and Pb-210 are reported for rain, snow, and hail samples taken at Argonne, Illinois, and Socorro, New Mexico. These natural radioisotopes are indicators of the sources of the aerosols contributing materials to the precipitation samples. The data presented indicate that the more soluble Be-7 is enriched in the precipitation samples with respect to Pb-210, as compared to the ratios of these radioisotopes found in aerosol samples. Use of the Pb-210/Pb-210 activity ratios as an internal clock indicated that the aerosols contributing to the precipitation ranged in age from 10 to 47 days. Levels of Be-7 ranged from 11 to 55 pCi L-1 for the samples, with the highest levels in a stratus precipitation event and in a thunderstorm with the lowest wet deposition rate. These results are discussed with regard to the potential for use of these radioisotopes in the determination of stratospheric-tropospheric mixing and in their geochemical usage as indicators of sedimentation rates.

The stochastic mixing model of cumulus clouds is extended to the case in which ice and precipitation form. A simple cloud microphysical model is adopted in which ice crystals and aggregates are carried along with the updraft, whereas raindrops, graupel, and hail are assumed to immediately fall out. The model is then applied to the 2 August 1984 case study of convection over the Magdalena Mountains of central New Mexico, with excellent results. The formation of ice and precipitation can explain the transition of this system from a cumulus congestus cloud to a thunderstorm.

A description of the known physical properties of a thunderstorm reveals that active cahrge separation occurs during that stage of the storm's life-cycle in which the growth of graupel by the accretion of supercooled droplets is the dominant process. Laboratory experiments under simulated thunderstorm conditions show that a graupel pellet, growing by the accretion of supercooled droplets, acquires negative charge as a result of collisions with ice crystals. Other experiments show that when two ice formations are placed in rubbing contact, the ice which is warmer, or which contains trace amounts of contaminants, acquires negative charge. Further experiments suggest that the charge separation results from potential differences which arise during the resolidification of a liquid layer formed at the ice-ice contact.

Calculations indicate that the graupel pellets in a thunderstorm, as a result of the acquisition of the latent heat of supercooled droplets, will achieve temperatures several degrees warmer than coexisting ice crystals. Thus the graupel pellets will acquire negative charge as a result of rubbing contacts with ice crystals. The graupel pellets have much higher fall velocities than ice crystals, thus accounting for the polarity of the main thunderstorm dipole. Measurements suggest that the amount of charge separated per graupel-crystal collision is adequate to account for the magnitude of the charges of the main dipole.

The magnitude and location of charge centers involved in lightning discharges can be determined from the value of the accompanying gradient change at seven points at the ground. Instrumentation, calibration, observation, and analysis techniques for thus locating thunderstorm charge-centers are described. The results of the analysis are presented.

The thunderstorm is found to be bipolar, with the negative center at a mean height of 25,000 ft msl or $-16\C$, and with the positive center located about 2000 ft above it. The occasional occurence of a positive center below the negative center is shown by the record of the leader processes of ground discharges and from other evidence.

Comparison is made with the work of Malan and Schonland in South Africa. While the results are in substantial agreement, important differences in the nature of streamer processes and the relative location of the negative centers tapped by successive elements of a single ground discharge are noted.

The cold environment found for many of the negative charge centers (as low as $-33\C$) strongly suggests that these centers might not have been produced by a glaze-ice mechanism. Riming and, perhaps, sublimation are probably the important processes for precipitation growth at such low temperatures.

It is shown, by use of Langmuir's precipitation growth data and Ludlam's hailstone heat economy data, that the glaze-ice thunderstorm charge generation mechanism suggested by E.J. Workman and the author is consistent with the thunderstorm cell development pattern observed by Workman and the author. The environment of charge separation is found theoretically. It also is shown that the initiation of precipitation in Midwestern thunderstorms without the involvement of ice crystals is consistent with Langmuir's cloud-droplet growth theory.

We describe the nanometer-sized phases in millimeter-sized spheres produced in a triggered lightning-strike experiment characterized by ultra-high heating of $\sim$ 10⁸ degrees s^-1 followed by similarly rapid quenching. The compositions of the aluminosilica glass spheres define a metastable Al₂O₃-SiO₂ eutectic at $\sim$ 40 wt. % Al₂O₃. The other phases are defined by the liquidus in the SiO₂-Al₂₃-Fe₃₄ (wt %) phase diagram whereby hercynite formed at $\sim$ 1750 degrees C and domains of Fe³⁺-rich cordierite glass that represent a ternary minimum melt at $\sim$ 1400 degrees C with cotectic glasses linking this glass and spinel. Metastable eutectics are potentially important to understand the phase relationships due to flash heating events.

A radio interferometer system is described which utilizes multiple baselines to determine the direction of lightning radiation sources with an angular resolution of a few degrees and with microsecond time resolution. An interactive graphics analysis procedure is used to remove fringe ambiguities from the data and to reveal the structure and development of lightning discharges inside the storm. Radiation source directions and electric field waveforms have been analyzed for different types of breakdown events for two lightning flashes. These include the initial breakdown and K type events of in-cloud activity, the leaders of initial and subsequent strokes to ground, and activity during and following return strokes. Radiation during the initial breakdown of one flash was found to consist of intermittent, localized bursts of radiation that were slow moving. Source motion within a given burst was unresolved by the interferometer but was detected from burst to burst, with negative charge being transported in the direction of the breakdown progression. Radiation during initial leaders to ground was similar but more intense and continuous and had a characteristic intensity waveform. Radiation from in-cloud K type events is essentially the same as for dart leaders; in both cases it is produced at the leading edge of a fast-moving negative streamer that propagates along a well-defined, often extensive, path. K type events are sometimes terminated by a fast field change that appears analogous to the field change of a return stroke. Dart leaders are sometimes observed to die out before reaching ground; these are termed "attempted leaders" and, except for their greater extent, are no different than K type events. Several modes of breakdown during and after return strokes have been documented and analyzed. One mode corresponds to the launching of a positive streamer away from the upper end of the leader channel, apparently as the return stroke reaches the leader start point. In another mode, the quenching of the dart leader radiation upon reaching ground reveals concurrent breakdown in the vicinity of the source region for the leader. In both instances the breakdown appears to establish channel extensions or branches that are followed by later activity of the flash. Finally, a new type of breakdown event has been identified whose electric field change and source development resemble those of an initial negative leader but which progresses horizontally through the storm. An example is shown which spawned a dart leader to ground.

A GPS-based system has been developed that accurately locates the sources of VHF radiation from lightning discharges in three spatial dimensions and time. The observations are found to reflect the basic charge structure of electrified storms. Observations have also been obtained of a distinct type of energetic discharge referred to as positive bipolar breakdown, recently identified as the source of trans-ionospheric pulse pairs (TIPPs) observed by satellites from space. The bipolar breakdown has been confirmed to occur between the main negative and upper positive charge regions of a storm and found to be the initial event of otherwise normal intracloud discharges. The latter is contrary to previous findings that the breakdown appeared to be temporally isolated from other lightning in a storm. Peak VHF radiation from the energetic discharges is observed to be typically 30 dB stronger than that from other lightning processes and to correspond to source power in excess of 100 kW; over a 6 MHz bandwidth centered at 63 MHz.

In July and August of 1989 the National Center for Atmospheric Research (NCAR) Sabreliner jet aircraft was used to probe electrically active and inactive convective storms over west central New Mexico to examine the production of odd nitrogen in the middle and upper troposphere by thunderstorms. In the anvil outflow or cloud top region of active and nonactive storms, the majority of flights showed that O-3 was reduced relative to the extracloud air owing to transport of ozone-poor air from lower altitudes. A similar result was found for active nitrogen (NOx) and total odd nitrogen (NOy) in nonelectrically active storms, but the reduction in NOy was also enhanced by removal of soluble constituents during convective transport. Examples of efficient removal from the gas phase are described. There was no evidence of O-3 production by lightning discharges. Indeed, O-3 was a goad tracer over the lifetime (similar to 1 hour) of the storms. During the active-to-mature stage of air mass thunderstorms, large enhancements in active nitrogen were observed in the anvil altitude region (9-11.8 km) and, in one case, in the midlevel outflow (near 7 km) of a dissipating thunderstorm. Two thunderstorms allow good estimates of the NOx production by lightning within or transport to the upper altitude region (8-11.8 km). Thunderstorms of August 12 and August 19 yield amounts in the range of 253-296 kg(N) and 263-305 kg(N), respectively. If, as an exercise, these amounts are extrapolated to the global scale on the basis of the number of cloud-to-ground and intracloud lightning flashes counted or estimated for each storm and a global flash frequency of 100 s(-1) the result is 2.4-2.7 and 2.0-2.2 Tg(N)/yr. Alternatively, an estimate for the two storms made on the basis of the average number of thunderstorms that occur per day globally (44,000) yields amounts in the range of 4.1-4.7 and 4.2-4.9 Tg(N)/yr, respectively. These estimates only apply to the production or transport of lightning generated NOx in or to the altitude region between 8 km and the top of the thunderstorm anvil (similar to 11.8 km in these studies). Since in some large-scale models, lightning-generated NOx is equally distributed by mass into each tropospheric layer, our estimates are roughly equivalent to those model runs that use a global source strength of about twice our estimate for the upper altitude region. In several flights where the region below the base of thunderstorms was examined, no large enhancements in odd nitrogen which could be clearly attributed to lightning were observed. Apparently, the aircraft was not in the right place at the right time. Thus no estimate of the NOx production by lightning that remains below similar to 8 km could be made.

Measurements of atmospheric electricity were made at a mountain-top observatory with some instrumentation at the earth's surface and with others carried aloft by captive balloons into the bases of thunderclouds overhead. The properties measured included the charge on individual precipitation particles, the electric current density carried by precipitation, the local electric field vector, and the electric conductivity of the air.

The electric fields aloft, near cloud base, were often twice the intensity of the surface fields and had large horizontal components that indicated non-uniform distributions of charge. The polar electrical conductivity of cloudy air was found to be about 2 x 10^-15 per $\Omega\cdot\m$ which is about 1/10 that of the clear air at the same altitude.

The polarity of charge on the precipitation aloft was almost invariably that of the local potential gradient; it did not show the well-known 'mirror-image' relation that is observed between the precipitation electricity at the ground and the local potential gradient during periods of point discharge. Use of our charge and conductivity data in Wormell's analysis of Wilson's ion capture mechanism suggests that ion capture in the subcloud region may explain adequately the precipitation charge arriving at the earth.

There was little correlation between the magnitudes of the precipitation charge aloft and the intensity of the electric field. We could not identify the origin of the charges found on precipitation aloft although one possible explanation is that they were derived from that residing on the cloud droplets from which the raindrops were formed.

Calculations of the electrical power transfer in the lower regions of developing thunderclouds by the fall of charged precipitation indicate local energy dissipation as predicted by Vonnegut, with rates of about 5 x 10^-6 W\cdot\m^-3.

The dual jet aircraft Sprites94 campaign yielded the first color imagery and unambiguously triangulated physical dimensions and heights of upper atmospheric optical emissions associated with thunderstorm systems. Low light level television images, in both color and in black and white (B/M), obtained during the campaign show that there are at least two distinctively different types of optical emissions spanning part or all of the distance between the anvil tops and the ionosphere. The first of these emissions, dubbed ''sprites'' after their elusive nature, are luminous structures of brief (< 16 ms) duration with a red main body that typically spans the altitude range 50-90 km, and possessing lateral dimensions of 5-30 km. Faint bluish tendrils often extend downward from the main body of sprites, occasionally appearing to reach cloud tops near 20 km. In this paper the principal characteristics of red sprites as observed during the Sprites94 campaign are described. The second distinctive type of emissions, ''blue jets,'' are described in a companion paper [Wescott el at, this issue].

Observations are presented in which the standard dual-polarization meteorological quantities (Z(DR), phi (dp), and rho (HV)) are determined from simultaneous horizontal (H) and vertical (V) transmissions. The return signals are measured in parallel H and V receiving channels. Because the parameters are determined from simultaneous measurements they are not affected by Doppler phase shifts that increase the variance of phi (dp) and rho (HV) when alternating H and V polarizations are transmitted. The approach has the additional advantage that a high-power polarization switch is not needed. The relative phases of the H and V components were such that the transmitted polarization was circular. Circular polarization is shown to detect horizontally oriented particles such as rain with the same effectiveness as linearly polarized transmissions, and optimally detects randomly oriented or shaped particles such as hail. Circular polarization also optimally senses nonhorizontally oriented particles such as electrically aligned ice crystals. By not needing to alternate between H and V transmissions it becomes practical to make polarization-diverse measurements by transmitting other orthogonal polarizations on successive pulses (e.g., left-hand circular and +45 degrees slant linear) to aid in identifying precipitation types. It is shown that rho (HV) from simultaneous transmissions provides the same information on randomly oriented scatterers as the linear depolarization ratio LDR from H or V transmissions, and that LDR does not need to be measured when information on particle canting is not important or is not needed.

During the summers of 1995 and 1996 we conducted broadband HF-UHF and narrowband VHF radio frequency (RF) observations of positive cloud-to-ground (+CG) flashes at Langmuir and Los Alamos laboratories, New Mexico. These observations indicate that positive leaders to ground produce no or very weak radiation from KF to UHF. The broadband system was able to record 2 ms data each time it was triggered. For a I-CG the system was usually triggered by the return stroke, and a 1 ms pretrigger period was coincident with the positive leader process. It was commonly observed that no or little radiation was associated with the leader process in the 1 ms pretrigger period. The narrowband VHF system employed a logarithmic power amplifier and recorded one 1 ms data each time it was triggered. The narrowband observations show that strong and often continuous radiation occurs at the beginning of the +CGs, but the radiation usually becomes intermittent and impulsive during the last few tens of milliseconds preceding the return strokes. The observations for most of the +CGs also show complete lack of radiation a few ms before the beginning of the return strokes, suggesting that the ongoing downward positive leaders were quiet at VHF, at least during the final few ms. The results of this study for natural positive leaders are in agreement with the results obtained from laboratory gap discharges and rocket-triggered lightning.

A detailed study of the complete sequence of VHF radiation events during intracloud (IC) flashes in Florida has shown that IC flashes often have a bilevel structure connected by a single upward channel. The lower and upper level channels appear to correspond to the main negative and upper positive charge regions of the storm, respectively. The IC flashes are characterized by active and final stages, each comprising about half of the overall flash duration. During the active stage, negative charge is transported upward in the cloud, first as a result of initial breakdown, which establishes the upward channel, and subsequently as a result of repeated breakdown events from the lower to the upper level. For the flashes of this study, the initial breakdown lasted 10 to 20 ms and propagated upward at a speed of 1.5 to 3 x 10⁵ m/s. It also established the beginning of one or more horizontal branches in the upper positive charge region, which were extended by the subsequent breakdown. Little or no radiation was detected along the upward channel during the subsequent breakdown, indicating that the channel remained more or less continuously conducting following the initial breakdown. After a time delay, the subsequent breakdown extended the lower-level channels in a retrograde manner horizontally away from the flash origin. At the end of the active stage the upward channel ceases to be conducting and the remaining, final stage of a flash is characterized by fast (10⁶ to 10⁷ m/s), well-defined "K"-type streamers that begin at successively greater distances along the lower-level channel and transport negative charge toward or into the base of the upward channel. The streamers are identical to those observed during the interstroke intervals and final stages of cloud-to-ground (CG) flashes. Typically, a few late-flash K streamers continue through the flash origin into and along an upper-level channel. The overall results are consistent with other observations of IC flashes and explain several features of the observations. Similar results are obtained for IC flashes in New Mexico storms. Comparison of the initial breakdown of IC and CG flashes shows that the IC breakdown is more intermittent than that of CG flashes, even though both have the same polarity and propagate at about the same average speed.

The authors use a numerical model of early electrification in thunderstorms, together with observations of a series of summer thunderstorms in New Mexico, to understand the roles of certain environmental factors in determining thunderstorm electrification. The results suggest that development of lightning depends sensitively on $\tau_{cz}$, the duration of significant updrafts in the charging zone between $-10\deg$ and $-25\C$. Model tests suggest that $\tau_{cz}$ is maximized for moderate cloud-base forcing and that $\tau_{cz}$ depends on environmental parameters in predictable ways. These results are used to investigate the relationships between lightning evolution and several functions of environmental quantities that have been suggested as lightning predictors.

Intense electric fields beneath thunderstorms produce electrical discharges (coronae) at the tips of trees, bushes and other sharp objects attached to the surface of the earth. We find typical corona current densities of about 1 nA m^-2 in an 8 kV/m field at the ground. The ions released into the air limit the magnitude of the field at the ground to about 10 kV/m. Our measurements beneath thunderstorms with a balloon-borne electric field meter show that the magnitude of the field a hundred metres above the ground is several times larger than at the ground; in one case the field 300 metres above the ground was 6 times that at the ground. The substantial thickness of the space charge layer and the speed with which it vanishes when the electric field strength declines imply that the charge carriers have subnstantial velocities (0.4 m/s) either because their mobilities are high or because they are carried by air motions.

Coronae also influence the time behaviour of the electric field at the ground. The field at the ground often changes very rapidly after a lightning flash. The rate of change decreases as the field approaches the value it had prior to the flash. In contrast, the field a hundred metres above the ground, which is often above most of the influence of space charge produced by coronae, increases more uniformly (linearly) during the time interval between lightning flashes. This behaviour is similar to that of the field farther aloft in the interior of the cloud. Our numerical simulations of the shapes of recovery curves indicate that the corona current density is more accurately described by a cubic function than by a quadratic function of the electric field strength at the ground.

Despite strong influences of coronae, three properties of the field at the ground accurately reflect what happens above the space charge layer. First, the rapid changes in electric field during a lightning flash are not usually affected by corona space charge. Second, when the field at the ground is nearly constant it usually has the same polarity as the field above the space charge layer. And third, when the field strength at the ground is nearly zero, and when certain other conditions are met, the time rate of change of the field at the ground is the same as that above the space charge layer.

The intense electric field beneath a thunderstorm often produces corona discharge from bushes and trees. The average steady state corona current per unit horizontal area, $J_c$, is approximately proportional to the time rate of change of the electric field at the ground when the value of the field changes sign following a lightning flash. This paper derives this relationship and proposes that the contribution of corona discharge to the charge budget of a thunderstorm be evaluated from measurements of the electric field at many sites below a thunderstorm rather than from direct measurement of the corona current.

High speed video of sprites show that they are typically initiated at an altitude of about 75 km and usually develop simultaneously upwards and downwards from the point of origin with an initial columniform shape. The initial development of sprites appears to be dominated by corona streamers with velocities in excess of 10(7) m/s. Many of the observed characteristics are consistent with a conventional breakdown mechanism for both sprite initiation and initial sprite development.

On August 14, 1998, three separate daytime sprite events were deteected via a unique extremely low frequency (ELF) sprite signature. The onset of the sprite ELF signatures was delayed by 11.0-13.2 ms from positive cloud-to-ground strokes which had attained exceptionally large charge moment (charge times height) changes of 3900-6100 C$/cdot$km. It is shown that a charge moment change of 6100 C$/dot$km may have been sufficient for conventional breakdown at $/sim$54 km altitude, assuming an experimentally measured ion conductivity profile of Holzworth et al., [1985]. The daytime sprites themselves contained unusually large charge moment changes of $/sim$2800 C$/dot$km, $/sim$1200 C$/dot$km, and $/sim$910 C$/dot$km.

The net electric charge density on all particles inside a thunderstorm can be estimated from measurements of the electric vector along a mainly vertical path of a balloon-borne sensor. In addition, the net charge density on precipitation particles such as raindrops, hail, and graupel can be deduced from measurements of the charge on each individual particle. This paper reports on two balloon soundings of electric field and precipitation charge in two different storms in central New Mexico. The two soundings are similar at lower heights and different at higher heights, possibly because one balloon ascended outside the main convective updraft core of the storm, while the other balloon ascended within the updraft. Six charge regions were inferred from the sounding outside the updraft core in the convective region. In the lowest four charge regions, including the lower positive, main negative, and two oppositely charged regions in between, the detected precipitation charge density (on particles with individual charges of at least 10 pC) was of the same polarity and as large or larger in magnitude than the: total charge density. At the top of and 0.4 km above the main negative charge region a large number density (about 125 m(-3)) of negatively charged precipitation particles were detected; these particles had larger mean charge (-48.6 pC) than any other group of positive or negative particles detected elsewhere in the sounding. The second electric field sounding was typical of updraft soundings below 8.0 kn, and the precipitation particles detected were all positively charged below 5.3 km and all negatively charged above. The detected precipitation charge density, although smaller in magnitude. was of the same polarity as the total charge density in the lower positive and main negative charge regions.

Three-dimensional lightning mapping observations have been used to estimate the peak source powers radiated by individual VHF events of lightning discharges. The peak powers vary form minimum locatable values of about 1 W typically up to 10-30 kW or more in the 60-66 MHz passband of the receivers. An energetic positive bipolar event radiated in excess of 300 kW peak power. The strongest radiation source tended to be observed in the upper part of storms, corresponding to the upper positive charge region, where the breakdown is of negative polarity. The results illustrate the bidirectional nature of intracloud discharges, with the largest source powers being along the negative portion of the discharge and an order of magnitude greater than the source powers along the positive portion. Overall, the source powers follow an approximate P^-1 distribution for powers above about 100 W. The radiation sources indicate the location of the main charge regions in a storm; sample comparisons with radar data show that the main negative charge coincided with the precipitation core.

Remarkable aspects of the thundercloud are its intense electrification, precipitation, and convection. A satisfactory understanding of how a thunderstorm works will require a continuing series of investigations to explore the complicated interrelationships among these phenomena. Until now the major effort has been devoted to studies of how precipitation causes electrification. For a century, investigations of thunderstorms have been dominated by the idea that lightning is produced by a charge-separation process within the cloud caused by falling precipitation. The origin of this idea, its implications, present status, and probable future are examined in the light of T. S. Kuhn's views on the nature of scientific progress. Despite some achievements, the results of research based on the precipitation theory have proved disappointing. For example, they have shed little light on important problems such as the factors that determine the polarity of the cloud electric dipole and the role that electricity plays in meteorological processes. During this century, with the discovery of cosmic rays and the ionization they produce in the air above the cloud, it has become apparent that other processes, which do not involve contact charge separation orf ailing precipitation, are also causing electrification. Thunderstorms exercise great influence, for both good and bad, on many human activities. In view of their great environmental importance, it is surprising how little is known about them and how little effort is being made to understand how they work. It is urged that the present limited thunderstorm research activities be expanded to include new, and possibly more productive, approaches.

A wire-tethered balloon system has been devised and used to measure directly the electrical potential of the atmosphere at altitudes up to 1 km above a mountain ridge in New Mexico.

Observations of summer thunderstorms developing over Mt. Withington, New Mexico, show that electrificatoin begins early in the cloud's development. Measurements within the cloud and above it show that electric charge accumulations similar in polarity to those of the mature storm begin to form before any echo can be seen with an x-band radar and before any electrical perturbations can be detected on the mountain summit beneath the cloud.

Measurements of potential gradient were made within the cloud with radiosondes supported on tethered balloons. It was found that the gradient within the cloud was far larger than that outside the cloud and that it reached values as high as 20 V/cm before the appearance of the radar echo. Measurements made from an airplane flying over the top of the growing cloud showed that here the fair weather potential gradient reversed before any radar echo could be seen. The development of the initial electrical activity appeared to be closely related to convective activity as indicated by tension on the line holding the tethered balloon and by the rapid rise of the cloud tops.

The initial radar precipitation echo within the cloud was frequently in the form of a hollow inverted cup that usually filled in and became solid in a few minutes.

An instrumented free balloon measured electric fields and field changes as it rose through a thundercloud above Langmuir Laboratory, New Mexico. The variation of the electric field with altitude implied that the cloud contained negative space charge of density $-0.6$ to $-4$ nC/m³ between 5.5 and 8.0 km MSL. The environmental temperature at these levels ranged from $-5\deg$ to $-20\C$. Our measurements imply that the areal extent of this negative charge center was significantly greater than that of the cloud's intense precipitation shafts. At altitudes grater than 8 km, the instrument ascended past net positive charge. We also inferred from our measurements positive space charge adjacent to the Earth's surface (concentration 0.6 nC/m³) and in the lower portion of the cloud (1.0 nC/m³). Electric field changes from intracloud lightning were interpreted by using a simple model for the developing streamer of the initial phase. Thunder source reconstructions provided estimates for the orientation of lightning channels. Seven `streamers' so analyzed propagated on the average, at 5 x 10⁴ m/s and carried a current of 390 A. The mean charge dissipated during a flash was 30 C.

A device for measuring the point discharge emitted from a 2m length of copper wire, using a radiosonde as carrier platform, has been developed in order to measure the distribution of charge regions inside thunderclouds with respect to temperature and altitude. No modifications to the radiosonde or in its standard receiving equipment are required.

In four penetrations of electrified clouds over central New Mexico, corona currents were induced in the conducting wire. The measurements were characterized by reversals in the polarity of the current which can be interpreted as a result of the balloon's rising past the level at which charge of one sign is concentrated.

With the use of an induction cylinder in conjunction with an optical array probe, simultaneous measurements were obtined of the size, shape, and charge of hydrometeors in a New Mexico thunderstorm. The particles with detectable charges (greater than 1--3 pC) were primarily graupel particles, and their charges are generally consistent with laboratory measurements of ice-ice collisional charging. The fraction with detectable charge increases with size, as does the magnitude of the mean charge. Also, the positively charged particles occur primarily at higher temperatures ($> -22\C$), and the degree of negative charging increases with decreasing temperature (down to $-35\C$).

The electric field of a thunderstorm is distorted by the metallic surfaces of airplanes and rockets. The amplitude of the local, distorted electric field can be measured at selected locations on the surface of the vehicle with electric field meters. Measuring the thunderstorm field entails finding the relations between the local fields at the meters and the ambient, undistorted electric field that would exist in the absence of the vehicle. Calibration can be performed in the following steps: (l) Find linear combinations $F_0, F_1, \cdots$, of local field amplitudes that are independent of the charge on the vehicle. (2) While the ambient electric field is constant in the earth coordinate system, rotate the vehicle (roll and turn or roll and pitch) and record the local field amplitudes and the roll, pitch, and heading angles. Find the direction of the electric vector and the ratios between all of the calibration coefficients that best fit the linear combinations of local field amplitudes. (3) Find the magnitude of one of the coefficients by comparison with a calibrated instrument; from this magnitude and all of the ratios, find all of the magnitudes. With this method of calibration, finding out how the linear combinations $F_0, F_1, \cdots$, depend on the ambient electri