TEN350
The waveform that hijacked the IEC lightning standards
CIGRE's 16 CAUTIONS
against using Berger's positive CG lightning data for engineering applications
CIGRE's 2013 Technical Brochure No. 549 contains the following cautions and arguments against relying on Berger's "positive CG lightning" data for engineering applications. (Last time we checked, such engineering applications would include recommending, designing or constructing lightning protection measures based on lightning waveforms inferred from Berger's data on positive CG lightning. )
According to CIGRE's Technical Brochure 549 the main problems with Berger's data on positive lightning include: insufficient data, statistically insignificant data, tainted samples, and misidentified lightning flash phenomena. Add them all together and you're left with the conclusion (stated 4 times in the report) that far more measurements need to be taken before any of Berger's data bearing on positive CG lightning waveform parameters can be corroborated. Until then, Berger's data is not statistically reliable enough for use in engineering applications. We have made these sixteen cautions more accessible by culling them directly from the report and including them in this web. Sincere thanks to CIGRE for allowing us to make the raw data available.
Below is provided the raw data and the section of the report in which the specific caution can be found--reachable by clicking on each line.
1) Re: Berger's data on positive lightning strokes, the data is very limited and may be influenced by the presence of the strike object on the mountain top. (Exec Summary)
2) Some of Berger's 26 "positive" flashes are likely to be not of the return-stroke type. You need to exercise caution in using Berger's waveshape parameters because the samples are even smaller than for his peak current distribution. Clearly additional data is necessary to establish statistically reliable distributions of positive CG lightning.(Exec Summary)
3) The positive current waveforms in Berger's data that were reported to have very large peaks (over 100kA) are likely to be the result of M-components (additional surges in the continuing current). (Sect. 2.3)
4) Berger's data between 10kA and 100kA are well supported by experimental data. Outside that range, the data is too small and the uncertainty is too large to determine a statistically significant distribution. (Sect. 3.1)
5) Re: Berger's positive lightning data: "particularly his extremely high (greater than 100kA or so) peak current tails require much larger sample sizes (probably of the order of thousands or more) than presently available (or to be available in the foreseeable future) to bring the uncertainties within an engineering accuracy range." (Sect. 3.1)
6) The sample of 26 directly positive lightning currents analyzed by Berger et al, commonly used as the primary reference in lightning protection, is apparently based on a mix of two different types of waveforms (a microsecond type waveform indicative of a single return stroke and a millisecond waveform (indicative of the additional surges or M-components in continuing currents.) (Sect. 7.4)
7) Some of Berger's 26 directly measured lightning flashes are likely NOT to be of the return-stroke type. Caution must be exercised, particularly for the waveshape parameters imputed to Berger's data, because the sample sizes are even smaller than for his peak current distributions. Clearly more data is needed for such waveshape parameters to be statistically significant.(Sect. 7.4)
8) The large charge transfers and action integrals observed in Berger's positive lightning is to be attributed to continuing currents. For simulation of worst case scenarios of positive lightning the report references Gamerota et al: Ipeak=350kA; time to decay to half-peak = 40 µs. (Sect. 7.4)
9) Some of Berger's 26 directly measured positive flashes are likely not to be of the return stroke type. Caution must be exercised particularly for the waveshape parameters because the sample size is too small to support a reliable waveshape. Clearly additional measurements for positive lightning return strokes are necessary to establish enough reliability for engineering purposes. (Sect. 7.5)
10) Berger's data on positive lightning is very scarce and is probably influenced by the fact that the strike object was on a mountain top. (Sect 10.2)
11) Berger's data on positive lightning is very limited and may be influenced by the tower being on the mountaintop. (Conclusions #2.)
12) Caution is to be exercised in trying to use Berger's data to imply waveshape parameters for positive lightning because the sample size is so small. Clearly, additional measurements of positive lightning return strokes are needed to establish reliable distributions of peak current and other parameters for this type of lightning." (Conclusions #8)
13) And let's not forget the most classic warning of all. The warning from Berger himself in Electra 41 where he said that his "positive strokes...do not have enough features to produce an acceptable mean current shape. This may also be due partly to the small number of positive strokes that were recorded.." (CIGRE Electra 41)
14) And lest we slight Anderson and Erikkson (Electra 69), remember they warned us that: "Berger has pointed out that all positive records (from Mt. San Salvatore) should, in fact, be classified as upward discharges." (CIGRE Electra 69)
15) Warning #14 is particularly important when read against this dire warning from the CIGRE 2013 TB 549 report: "It is very important in any comparison of CG lightning parameters to exclude any upward lightning from the analysis." (Sect. 9.1)
16) The last warning from the CIGRE 2013 report is the reference to Carlos Romero et. al who in 2010 and 2011 recorded 30 positive lightning flashes on the Säntis Tower, also in the Swiss Alps, and less than 200km from Berger's Mt. St. Salvatore. All of Romero's 30 positive lightning flashes were upward lightning. Every one. Romero's results lend credence and probability to warnings 13 and 14 above. (Sect. 8.6)
