EGU General Assembly 2020
© Author(s) 2020. This work is distributed under
the Creative Commons Attribution 4.0 License.

Radio emission from fast streamers

Nikolai Lehtinen
Nikolai Lehtinen
  • University of Bergen, BCSS, Institutt for fysikk og teknologi, Bergen, Norway (

A new computational approach, based on treating an electric streamer as a nonlinear instability, allows to determine unambiguously its parameters, for a given streamer length and external electric field, which may be nonuniform. Among the determined parameters are the speed, current and conductivity inside the streamer. These parameters may vary over orders of magnitude, depending on external conditions.

We use these parameters to calculate the radio emissions which would be observed on the ground from fast discharges produced in lightning, in which streamer velocities approach a significant fraction of the speed of light. Fast discharges play an important role in lightning initiation and may be responsible for production of Terrestrial Gamma Flashes (TGF). They manifest themselves in ground-based radio observations as Narrow Bipolar Events (NBE), to which the calculation results are compared.

We will discuss conditions, the effect of which on streamer propagation (and therefore electromagnetic radiation) may be quantified with the used computational method. These include (i) the external electric field modification due to charges deposited by previous streamers; and (ii) electron attachment inside the streamer channel, which is strongly affected by cloud humidity.

How to cite: Lehtinen, N.: Radio emission from fast streamers, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-9912,, 2020

Comments on the presentation

AC: Author Comment | CC: Community Comment | Report abuse

Presentation version 2 – uploaded on 04 May 2020
cosmetic changes
  • AC1: A quick go-through (will be repeated during the chat), Nikolai Lehtinen, 04 May 2020

    We use a new computational approach ("parametric model of streamers") to calculate unambiguously streamer parameters (described in slide 4, with more details available at Application of this approach to streamers at sea-level air gave reasonable (∼ several 10%) agreement with experiment (for velocity and negative streamer threshold measurements) and hydrodynamic modeling in which several codes were tested [Bagheri et al., 2018].

    This approach is applied to long streamers in uniform electric fields, which are shown to accelerate to speeds close to the speed of light ("fast streamers" of Rison et al. [2016]). 

    Long (fast) negative streamers may propagate at fields lower than laboratory-measured threshold fields (0.75-1.25 MV/m), because the threshold depends on length L and decreases with increasing length (slide 8)

    The fast streamers develop high speeds and currents (slides 9-10). Despite high current, they do not heat to form leaders because the propagation time is very short (slide 11).

    The propagation is stopped due to the field of the charge they leave behind (slide 12), with the configuration of the field (very approximate!) described in slide 14.

    Electromagnetic pulses from this process are shown in slides 15-16 (for positive and negative streamers, though they look very similar for similar fields). These pulses are the observed Narrow Bipolar Events (NBE).

  • CC1: Comment on EGU2020-9912, Alexander Kostinskiy, 05 May 2020

    Dear Nicholas,

    You have proposed an interesting theoretical model of the movement of streamers. You may be interested in comparing it with experimental data on the physics of a long spark.

    Slides 8-12. Based on what experiments do you take an electric field that supports the movement of negative streamers less than 10 kV / (cm atm)? With an electric field of 6-9.5 kV/(cm atm), long negative streamers will definitely never move (I mean a length of 0.5-3 meters). These are repeatedly and well-verified data and some experiments give a minimum field of 12 kV/(cm atm).

    Slide 9. Unfortunately, in experiments at pressures of 0.3-1 atm, positive streamers 1-3 meters long never move at speeds even of 10^7 m/s (such high speeds were measured only at fields of about 150 kV/(cm atm)) . With your electric fields, the speed of positive streamers does not exceed 2-3*10^6 m/s (even at 10 kV/cm). About negative streamers I wrote above. The experiments of Rison et al. 2016 (slide 6) can hardly be called experiments on measuring the speeds of streamers, since the movement of their centroids is calculated by complex processing of radio signals and cannot be unambiguously interpreted. Moreover, FPB by their mathematical model is a giant volumetric streamer flash, and the centroid of the interferometer is basically one point of the maximum signal in the processing window. The centroid can also be a phase wave of statistical fluctuations inside the FPB and a phase wave of ignition of streamer flares. Therefore, it cannot be ruled out that FPB may be just another physical process. In addition, with a length, the speed of the streamers, even in an increasing electric field of 6-8 kV/(cm atm), grows weakly. Streamer flash speeds were measured by streak cameras with image amplification and it is difficult to distrust these experiments. I can send you links to these experiments. Perhaps the experiments are different from your calculations due to the fact that the streamer flash is always triggered, and not just one streamer. There are hundreds of streamers in a regular streamer flash. In this case, the charges of the streamer heads reduce the electric field on other heads. As you know, Bazelyan-Raizer wrote about this in a book about a long spark.



    • AC2: Reply to CC1, Nikolai Lehtinen, 05 May 2020

      Dear Alexander,

      Thank you very much for the detailed comment!

      1. For laboratory negative streamers, the threshold is indeed higher. This is mentioned in slide 8 here, but calculated in detail in Figure 9 in . The theory predicts length-dependent thresholds, for negative streamers it is decreasing with length. (In experiment [N. L. Allen and P. N. Mikropoulos, J. Phys. D 32, 913(1999).], the length dependence of a positive streamer threshold was measured, the threshold was increasing with length.) Thus, a negative streamer has to be long to propagate in lower field. Or, a non-uniform field is needed, some preliminary calculations were presented in my AGU poster [, last column, section "Electrode Field"].

      2. The velocity of streamer is also length-dependent (in uniform fields!), as shown in slide 9 here for long streamers, but for short streemers see Figure 6a in , where indeed the streamers only reach 10^7 m/s at 1.5 MV/m, but for maximum lengths of about 20 cm. In the long sparks, the fields are very non-uniform, which also makes the situation different. For a simple preliminary model of a single streamer, see the above link to my AGU poster.

      The model does not have branching and leaders, and it would be fascinating to see a theory which would explain two types of negative streamers observed by Gorin et al [1976], creation of pilot systems and leaders.

      I have studied your hypothesis on creation of FPB which involves unusual plasma formations (UPF). As I understand, this situation differs from laboratory conditions in that there are random pockets of space charge. Our theory does not include them yet, it would be interesting to include the space charges, too.

      Thank you for discussion,
       - Nikolai Lehtinen

    • AC3: Reply to CC1, Nikolai Lehtinen, 05 May 2020

      Dear Alexander,

      Another thing that we have not included yet in these preliminary calculations are the processes in the streamer channel. The removal of electrons due to attachment and recombination is responsible for the positive streamer threshold, and may indeed reduce the streamer speed. This is still to be done for the long streamers, but for the short streamers some preliminary results are presented in my AGU poster [, last column, section "Role of attachment and positive streamer threshold"].

      Best regards,
       - Nikolai.

  • CC2: Comment on EGU2020-9912, Brian Hare, 05 May 2020

    these results seem to be in a range close to what we think we see with LOFAR (see D1866). Do you think it is possible to compare your model with our observations?

    • AC4: Reply to CC2, Nikolai Lehtinen, 05 May 2020

      Dear Brian -

      Yes, definitely. Although, there is still a free parameter in the model - the radius of the area from which the charge is collected, R_Q (=10 m in the presentation).

      Best regards,

       - Nikolai.

Presentation version 1 – uploaded on 04 May 2020 , no comments