The large-amplitude radiation field pulses produced by intracloud lightning discharge processes have been recorded with submicrosecond time resolution. The wave forms are distinctly different from those produced by return strokes in cloud-to-ground lightning, yet they are surprisingly alike within a discharge and in different discharges. The shapes tend to be bipolar, with two or three narrow, fast-rising pulses superimposed on the initial half cycle. Pulses with a positive initial polarity are usually produced in the several tens of milliseconds prior to the first return stroke in a cloud-to-ground discharge. Positive pulses tend to occur at regular intervals and have a mean full width of about 40 + 13 us. Negative pulses are usually produced during isolated cloud discharges at more random intervals and have shapes similar to the positive pulses but with more variability. The implications of the field shapes and polarities for the physics of intracloud discharge processes are discussed.

We present a characterization of Florida lightning return stroke electric and magnetic fields derived from simultaneous measurements of the fields at two separate stations, one station being within 15 km of the lightning, the other at either about 50 or 200 km from the lightning. We give (1) examples of correlated wave forms, (2) typical first and subsequent stroke wave forms over the distance range 1.0–200 km, and (3) the following statistical data from which the typical wave forms were derived: for electric field, rise time, initial peak value, ramp starting time, ramp slope, value at 170 μs, ratio of value at 170 μs to initial peak, zero‐crossing time for 50 and 200 km wave forms; for magnetic fields, time of hump following initial peak, ratio of hump value to initial peak value, zero‐crossing time for 50 and 200 km wave forms. Return stroke electric and magnetic field characteristics appear to be independent of location in Florida.

Lightning discharges to ground produce radiation fields which have structures that are surprisingly similar for different return strokes in the same discharge and also for different discharges. The first stroke begins with an initial portion or front which rises slowly for 2-8 •ts to about half of the peak field amplitude. Return strokes subsequent to the first have initial fronts which last 0.5-1.0 sts, on the average, and which rise to about 20% of the peak field. Following the front, both first and subsequent stroke wave forms rise abruptly to peak with 10-90% rise times of 0.2 •ts or less when the field propagation is over seawater. After the initial peak most strokes have a small second peak or shoulder within about 2 •ts, and all first strokes have several large subsidiary peaks at intervals of 10-30 •ts following the first peak. The physical processes which produce these structures are not well understood, but the data suggest that there may be rather large currents in the upward connecting discharges that occur just prior to first strokes and that all return stroke currents contain a large submicrosecond component. The small second peak may be produced by a current oscillation or by a reflection of a traveling wave in the return stroke just after its onset. The large subsidiary peaks in first strokes are probably produced by the effects of branches. The similarity of subsequent stroke fields within a flash suggests that the currents and velocities of different subsequent strokes in the same discharge are often very nearly the same. At the present time most of our knowledge about the development of lightning return strokes is based on analyses of a limited number of time-resolved photographs and direct measurements of the currents produced by strikes to instrumented towers [Uman, 1969]. Recently, several workers have been attempting to infer characteristics of return stroke velocities and currents from remote measurements of the electromagnetic fields produced by lightning within about 200 km [Urnan and McLain, 1970b; Urnan et al., 1973a, b; Linet al., 1976; Leise and Taylor, 1977; Price and Pierce, 1977]. This approach was originally pioneered by Norinder and co-workers in Sweden [Norinder, 1928[Norinder, , 1956Norinder and Dahle, 1945] and more recently by Wagner [1960] in the United States. If this method proves to be viable, it will enable the characteristics of many discharges to be studied in different geographical locations and under a variety of meteorological conditions. The shapes of the electric and magnetic fields produced by the stepped and dart-stepped leader processes which immediately precede return strokes in lightning discharges to ground have been discussed in two previous papers [Krider and Radda, 1975;Krider et al., 1977]. As the stepped leader nears the ground, the electric field at the surface becomes very large, and usually one or more upward propagating discharges are initiated which rise and join the leader channel at a point a few tens of meters above the surface (see Uman [1969...

The electric fields produced by stepped and dart-stepped leaders which immediately precede return strokes in lightning discharges to ground have been recorded in Florida and Arizona. The mean interval time between normal steps is about 16 us, and the mean interval between dart steps is 6-8 us. The amplitudes of leader pulses in Florida increase just prior to the return stroke, the largest usdally being about 10% of the return stroke peak. In Arizona the leader pulse amplitudes are smaller than those in Florida, in relation to the return stroke, and are not as easy to identify. The shapes of the fields produced by normal steps are similar to dart steps, and the dart steps are very similar to regular sequences of pulses produced by many intracloud discharges. The 10-90% rise times of individual step wave forms are often less than 0.3 us, and the full width at half maximum of a step pulse is typically 0.4-0.5 us under conditions where the propagation distortion is minimal. The amplitudes and the shapes of leader step wave forms suggest that the peak step current is at least 2000-8000 A close to the ground and that the maximum rate of change of step current is 6-24 kA/tts or larger. A rough estimate of the minimum charge lowered during the formation of a step is (1-4) X 10 -3 C.

Forty simultaneous, submicrosecond time‐resolved measurements of triggered lightning returnstroke current (I), current derivative (dI/dt), and electric field derivative (dE/dt) were made in Florida and France in 1985 and 1986. Peak currents ranged from about 5 to 50 kA, peak dI/dt amplitudes from 60 to 260 kA/μs in 1985 and from 20 to 140 kA/μs in 1986. The mean peak dI/dt values, 111 kA/μs (1985) and 68 kA/μs (1986), are 2–3 times higher than data from instrumented towers, and peak I and dI/dt appear to be positively correlated. The dE/dt and dI/dt waveform pairs have similar shapes, and the peak amplitudes are linearly proportional. Return‐stroke velocity, computed from the ratio of peak dE/dt and dI/dt signal amplitudes using an expression derived from the radiation field term of the transmission line model (TLM), averaged 2.9×108 m/s and 3.0×108 m/s in 1985 and 1986, respectively, which is about 2 times higher than most optical measurements. The TLM velocity may be erroneous because (1) the dE/dt measurement was made only 50 m from the lightning channel, where fields other than the radiation field component, that is near fields, may contribute to the total dE/dt and (2) fine structure on the measured E fields was not consistent with a single upwardly propagating return‐stroke current wave assumed by the TLM (two waves are more consistent).