| Single-station electric field (E) and electric field derivative (dE/dt) waveforms were recorded at digitization rates up to 400 MS/s during 1989 and 1990 at the Kennedy Space Center. Narrow bipolar pulses (NBP's) were found to occur separate from typical lightning events, but to be thunderstorm related. Frequency spectra for E obtained from NBP's dropped at a rate of close to 1/f up to 125 MHz. In comparison, the frequency spectra reported by Willett et al. (1989) for similar pulses dropped as 1/f up to about 20 MHz and became flat afterwards up to their 50 MHz Nyquist frequency. NBP's contained higher energy than return strokes above 10-20 MHz. NBP Spectra found from the output of digitally simulated narrowband receivers tended to underestimate the wideband frequency spectra by as much as 10 dB, indicating that the spectra obtained using narrowband receivers are unreliable.; Initial E-field peaks of NBP's had a mean rise time of 1.38 {dollar}mu{dollar}s. Large positive dE/dt pulses had a mean half width of 7 ns, much shorter than the 49 ns reported by Willett et al. (1989).; The existence of a single process responsible for VHF radiation from lightning, such as that proposed by Labaune et al. (1990), was tested using deconvolution methods on the NBP waveforms. Our analysis failed to identify a single basic component in these pulses.; Electric fields from lightning strikes at distances within 1-2 km consistently exhibited a chaotic behavior during the stepped leader, whereas distant stepped leaders did not. This "chaos" ranged from pulses occurring at rates close to one pulse per {dollar}mu{dollar}s to a continuous noise-like high frequency signal with frequency components extending beyond 120 MHz.; In agreement with other reports in the literature, we found that HF radiation following return strokes peaked 20-30 {dollar}mu{dollar}s after the onset of the return stroke, and persisted for several tens of microseconds after the peak. However, the short propagation path (less than 7.5 km) over salt water does not support the widely accepted hypothesis that the delayed peak arises as a result of propagation effects. |
