Predicting lightning sourced electromagnetic fields

Abstract

Sferics used in audio-frequency magnetotellurics (AMT) for shallow conductivity-depth soundings (up to 400 m in a 1000 Ωm half-space at 1.5 kHz) are detected by global lightning networks, which catalogue the time and location of up to four million lightning strikes per day. Passive AMT surveys do not use source information, but rather assume that global lightning activity provides sufficient signal if data are acquired for long enough. This assumption not only limits survey productivity, but can also result in poor signal-to-noise ratios (S/N), particularly at dead-band frequencies (1.5–5 kHz). To increase S/N and survey productivity, we used lightning network data as a proxy for direct measurement of sferic source information. We carried out a Global Positioning System (GPS) synchronised AMT survey, and for each lightning strike in the catalogue we modelled earth-ionosphere waveguide propagation to accurately predict arrival time for every predicted sferic in our time series data. We extracted a window of data around each predicted sferic and stored these time series into a structured database with associated lightning network data, such as lightning peak current, polarity and geographical coordinates. Two examples from our sferic database show that lightning network data are accurate enough to predict amplitude and polarisation of sferics at our survey site. For large lightning peak current, we observed a strong correlation between source proximity and increased S/N, particularly at dead-band frequencies, which often correlate to induction scales pertinent to mineral exploration. Our findings establish an interconnected relationship between lightning network data and lightning sourced fields, introducing known source AMT.