AbstractIn upward flashes, charge transfer to ground largely takes place during the initial continuous current (ICC) and its superimposed pulses (ICC pulses). ICC pulses can be associated with either M‐component or leader/return‐stroke‐like modes of charge transfer to ground. In the latter case, the downward leader/return stroke process is believed to take place in a decayed branch or a newly created channel connected to the ICC‐carrying channel at relatively short distance from the tower top, resulting in the so‐called mixed mode of charge transfer to ground. In this paper, we study the electromagnetic fields associated with the M‐component charge transfer mode using simultaneous records of electric fields and currents associated with upward flashes initiated from the Säntis Tower. The effect of the mountainous terrain on the propagation of electromagnetic fields associated with the M‐component charge transfer mode (including classical M‐component pulses and M‐component‐type pulses superimposed on the initial continuous current) is analyzed and compared with its effect on the fields associated with the return stroke (occurring after the extinction of the ICC) and mixed charge transfer modes. For the analysis, we use a 2‐Dimentional Finite‐Difference Time Domain method, in which the M‐component is modeled by the superposition of a downward current wave and an upward current wave resulting from the reflection at the bottom of the lightning channel (Rakov et al., 1995, https://doi.org/10.1029/95JD01924 model) and the return stroke and mixed mode are modeled adopting the MTLE (Modified Transmission Line with Exponential Current Decay with Height) model. The finite ground conductivity and the mountainous propagation terrain between the Säntis Tower and the field sensor located 15 km away at Herisau are taken into account. The effects of the mountainous path on the electromagnetic fields are examined for classical M‐component and M‐component‐type ICC pulses. Use is made of the propagation factors defined as the ratio of the electric or magnetic field peak evaluated along the mountainous terrain to the field peak evaluated for a flat terrain. The velocity of the M‐component pulse is found to have a significant effect on the risetime of the electromagnetic fields. A faster traveling wave speed results in larger peaks for the magnetic field. However, the peak of the electric field appears to be insensitive to the M‐component wave speed. This can be explained by the fact that at 15 km, the electric field is still dominated by the static component, which mainly depends on the overall transferred charge. The contribution of the radiation component to the M‐component fields at 100 km accounts for about 77% of the peak electric field and 81% of the peak magnetic field, considerably lower compared to the contribution of the radiation component to the return stroke fields at the same distance. The simulation results show that neither the electric nor the magnetic field propagation factors are very sensitive to the risetimes of the current pulses. However, the results indicate a high variability of the propagation factors as a function of the branch‐to‐channel junction point height. For junction point heights of about 1 km, the propagation factors reach a value of about 1.6 for the E‐field and 1.9 for the H‐field. For a junction height greater than 6 km, the E‐field factor becomes slightly lower than 1. The obtained results are consistent with the findings of Li, Azadifar, Rachidi, Rubinstein, Paolone, et al. (2016, https://doi.org/10.1109/TEMC.2015.2483018) in which an electric field propagation factor of 1.8 was inferred for return strokes and mixed‐mode pulses, considering that junction points lower than 1 km or so would result in a mixed mode of charge transfer, in which a downward leader/return‐stroke‐like process is believed to take place. It is also found that the field enhancement (propagation factor) for return stroke mode is higher for larger ground conductivities. Furthermore, the enhancement effect tends to decrease with increasing current risetime, except for very short risetimes (less than 2.5 μs or so) for which the tendency reverses. Finally, model‐predicted fields associated with different charge transfer modes, namely, return stroke, mixed‐mode, classical M‐component, and M‐component‐type ICC pulse are compared with experimental observations at the Säntis Tower. It is found that the vertical electric field waveforms computed considering the mountainous terrain are in very good agreement with the observed data. The adopted parameters of the models that provide the best match with the measured field waveforms were consistent with observations. The values for the current decay height constant adopted in the return stroke and mixed‐mode models (1.0 km for the return stroke and 0.8 km for the mixed‐mode pulse) are lower than the value of 2.0 km typically used in the literature.