REAS3: A revised implementation of the geosynchrotron model for radio emission from air showers

Over the past years, the freely available Monte Carlo-code REAS which simulates radio emission from air showers based on the geosynchrotron model, was used regularly for comparisons with data. However, it emerged that in the previous version of the code, emission due to the variation of the number of charged particles within an air shower was not taken into account. In the following article, we show the implementation of these emission contributions in REAS3 by the inclusion of ``end-point contributions’’ and discuss the changes on the predictions of REAS obtained by this revision. The basis for describing radiation processes is an universal description which is gained by the use of the end-point formulation. Hence, not only pure geomagnetic radiation is simulated with REAS3 but also radiation due to the variation of the net charge excess in the air shower, independent of the Earth’s magnetic field. Furthermore, we present a comparison of lateral distributions of LOPES data with REAS3-simulated distributions. The comparison shows a good argeement between both, data and REAS3 simulations.

💡 Research Summary

**
The paper presents REAS3, a major revision of the REAS Monte‑Carlo code used to simulate radio emission from extensive air showers (EAS). The authors identify a critical shortcoming in the previous version, REAS2: it ignored radiation arising from the time‑varying number of charged particles (the so‑called charge‑excess effect). REAS2 only accounted for radiation generated along particle trajectories due to geomagnetic deflection, neglecting the abrupt acceleration and deceleration that occur when particles are created or annihilated.

To remedy this, the authors adopt the “end‑point formalism,” a universal description of electromagnetic radiation that treats particle tracks as a series of straight‑line segments connected by instantaneous kinks. At each kink the particle velocity changes abruptly, and the time‑integrated electric field can be expressed analytically (see Eq. 1 in the paper). The start and end of a track are special kinks where one of the velocities is zero, thus naturally incorporating radiation from particle creation and disappearance. This approach simultaneously captures geomagnetic (v × B) radiation and the charge‑excess contribution without the need for separate modules.

Simulation results demonstrate three key improvements over REAS2. First, the radio pulse shape changes from a unipolar to a bipolar form, reflecting the physical expectation that radiation is emitted only over a finite time interval. Second, the frequency spectra now drop to zero at zero frequency, eliminating the unrealistic low‑frequency plateau seen in REAS2 and bringing the spectra into better agreement with the macroscopic MGMR model. Third, because the charge‑excess component does not depend on the Earth’s magnetic field, an azimuthal asymmetry appears: the east‑west signal is stronger than the west‑east signal, a feature absent in REAS2.

The authors validate REAS3 against measurements from the LOPES radio‑detector array. For each LOPES event they generate 200 CONEX showers, select a “typical” shower with an Xmax close to the average, and run both REAS2 and REAS3 using the same CORSIKA particle data. The comparison shows that REAS3 reproduces the lateral distribution of the radio signal far more accurately. In particular, REAS3 yields flatter lateral slopes, matching the subset of LOPES events that exhibit unusually flat distributions—something REAS2 could not achieve. Although a few events still show discrepancies, the overall agreement is markedly better, confirming that the inclusion of the charge‑excess term is essential for realistic modelling.

The paper also discusses the physical origin of the charge‑excess radiation: the net negative charge in an air shower varies as the shower develops, producing a radially polarized electric field component even in the absence of a magnetic field. Simulations with B = 0 G confirm this effect, showing a purely radial pattern with a magnitude about one‑third of the total field at 60 MHz for observer distances up to 200 m.

In conclusion, REAS3 provides a self‑consistent, parameter‑free, time‑domain simulation that fully incorporates both geomagnetic and charge‑excess mechanisms, based on the detailed particle information supplied by CORSIKA. The code is made freely available, paving the way for broader adoption and further refinements such as full detector response modelling. This work represents a significant step toward reliable theoretical predictions for radio detection of cosmic‑ray air showers.