Discussion

This paper presents two main points: the LASA/ NLDN comparison for the determination of the LASA accuracy, and the intial LASA/LDAR comparison in order to understand the low event coincidence rate between NLDN and FORTE.

The theoretical 600 m location accuracy was postulated under conditions of favorable geometry and assuming an excellent waveform cross correlation capability. Despite the favorable geometry within the 1998 New Mexico array, as illustrated in Fig. 6, the optimal predicted accuracy of 600 m was not achieved. This may be attributed to several factors: the first is the fact that the NLDN lightning locations do not necessarily represent the true source locations. The uncertainty for the standard NLDN data product is 0.5 km. It is not known whether this accuracy is achieved for the region of New Mexico that includes the 1998 LASA stations. The effect of NLDN `loosened criteria' (explained earlier) is also not known. A second consideration is that LASA waveform cross correlations are not perfect. With identical waveforms at all stations it would be possible to determine the actual DTOAs to within 1 us. Propagation over the finitely conducting ground, ionospheric reflections, and static-near and inductive-intermediate field influences all affect the wave shapes. Some of these effects are illustrated by the waveforms in Fig. 2. A third source of contamination is from incidental coincidences. The events used for this comparative study were selected by finding +/- 100 us LASA/NLDN coincidences. Event ranges from the NM array centroid were based on the LASA event locations. Incidental NLDN events that occurred within the +/- 100 us window with LASA events, but occurred at great distances could be included in Fig. 6. Within the sub-arrays, the LASA geolocation error is comparable to the reported NLDN errors.

A surprising result of the FORTE/NLDN comparison is a coincident event detection rate of only ~1% (personal comm., A. Jacobson). In order to understand the low FORTE/NLDN coincidence, further VLF/VHF comparisons are currently underway, using LASA (VLF) and co-located ground VHF systems in order to collect more VLF information than provided by NLDN (waveforms, rather than waveform parameters) and also to increase the number of VLF/VHF coincidences (compared to FORTE/LASA). The LASA/LDAR results presented illustrates one of the most difficult aspects of VLF/VHF comparisons-namely, the large number of VHF events reported for each VLF event (~100 for the LASA/LDAR comparison). One solution taken by other investigators comparing LDAR with other systems is to group the LDAR events and do a climatological study rather than an event-based study [Boccippio et al., 2000b]. Currently we are examining methods of making an event-based comparison between LASA and VHF systems (LDAR and LMA).