85th AMS Annual Meeting

Tuesday, 11 January 2005
Carlos Morales, University of Sao Paolo, Sao Paulo, 05508-900, Sao Paulo, Brazil; and O. Pinto Jr. and E. N. Anagnostou
This paper evaluates the location error accuracy and the detection efficiency of two existing global VLF lightning detection networks, named TOGA and ZEUS, over Brazil based on a comparison with the Brazilian lightning detection network, RINDAT. In order to make these intercomparisons, we have constrained the intercomparison over the South and Southeast Brazil, where we expect to have 80-90% detection efficiency and 0.5-2 km location accuracy for RINDAT, which will be used as our “ground truth”. To compute the location error, we seek coincident measurement in a time window of 0.0001 seconds, mainly because each system measures in different frequency and consequently they can observe distinct lightning process, and also to account for timing errors in the solution. For the detection efficiency, we binned all the lightning strokes from each network in 2x2 degrees (latitude and longitude) and in one-hour interval, and we make sure that each network was operating in that time. The evaluation of these errors was performed during November of 2003, where have available RINDAT, TOGA and ZEUS data. The results computed until the present moment show the following. 1) Location error: By applying the area and time constrain we were able to find 2370 coincident measurements of TOGA and RINDAT, and 2020 ZEUS and RINDAT. For this matches we have found that TOGA mean location error (standard deviation) is 14.64 km (20.17) while for ZEUS we found that is 62.84 km (21.49); 2) Detection efficiency: Since, there is a strong time dependency of VLF signal propagation during day and night, we have computed the mean detection efficiency in both periods. The preliminary results show that TOGA has a nighttime (daytime) detection efficiency of 1.56 (0.62) while ZEUS has 25.44 (2.25). The final version of this study will show the spatial and temporal distribution of location errors and its detection efficiency. In addition to this analysis, an effort will be make to depict the errors over South America by an intercomparison with LIS. Therefore, these two existing lightning networks can complement the lightning observations of the Brazilian lightning detection network over the Amazon and Northeast of Brazil.

Supplementary URL: