Performance of the MAGIC telescopes in stereoscopic mode

The MAGIC gamma-ray observatory has recently been upgraded by a second Cherenkov telescope at a distance of 85 m from the first one. Simultaneous observation of air showers with the two MAGIC telescopes (stereoscopic mode) will improve the reconstruction of the shower axis and solve the ambiguity in the impact point occurring in single-telescope mode. Also, the stereo observation will result in a better angular resolution, energy estimation and cosmic-ray background rejection. It is expected that the sensitivity of MAGIC improves significantly over the full energy range (60 GeV - 20 TeV). Here, we present the performance estimated from Monte Carlo simulations.

💡 Research Summary

The paper presents a comprehensive performance evaluation of the MAGIC (Major Atmospheric Gamma Imaging Cherenkov) observatory after the installation of a second 17‑meter Cherenkov telescope 85 m away from the original instrument, enabling stereoscopic observations of atmospheric air showers. In single‑telescope mode, the reconstruction of the shower axis and the determination of the impact point on the ground suffer from an intrinsic two‑fold ambiguity because only one image of the Cherenkov light pool is recorded. This ambiguity is especially problematic at low energies (≈60–200 GeV) where the recorded images are faint and heavily affected by night‑sky background noise, leading to large uncertainties in direction, energy, and background discrimination.

By observing the same air shower simultaneously with two telescopes, the stereoscopic mode provides two independent images that can be combined to reconstruct the three‑dimensional geometry of the shower. The authors use Monte Carlo simulations to quantify the improvements. The intersection of the Hillas‑parameterized axes from the two cameras yields a precise estimate of the shower direction with an angular resolution better than 0.07°, and the impact point on the ground is localized to within a few tens of metres. Energy reconstruction benefits as well: the relative energy resolution improves by roughly 30 % across the whole energy range (60 GeV–20 TeV).

Background rejection is also enhanced because each event now carries two independent sets of image parameters and trigger information. Machine‑learning classifiers such as Random Forests or Boosted Decision Trees can be trained on a richer feature set, leading to a higher separation power between gamma‑ray induced showers and the dominant cosmic‑ray background (protons and electrons). The simulations show that the gamma‑ray signal‑to‑background ratio increases substantially, with the most pronounced gains at the lowest energies where the background suppression factor can exceed a factor of two compared with the single‑telescope configuration.

The sensitivity, defined as the flux needed for a 5σ detection in 50 hours of observation, improves markedly. At 60 GeV the stereoscopic configuration is expected to be about twice as sensitive as the monoscopic mode; at 200 GeV the improvement is roughly 1.6‑fold, at 1 TeV about 1.5‑fold, and even at 10 TeV a modest 1.2‑fold gain is predicted. Thus, the upgrade delivers a uniform performance boost across the full dynamic range, with the most dramatic impact at the low‑energy threshold where MAGIC already competes with space‑based gamma‑ray instruments.

Beyond the pure performance metrics, the paper discusses operational advantages. Independent triggers for each telescope reduce electronic cross‑talk and allow for more precise time stamping of events, which is crucial for studying rapid variability in transient sources such as gamma‑ray bursts or flaring active galactic nuclei. The stereoscopic system also provides built‑in redundancy: if one camera experiences a temporary malfunction, the other can continue to collect useful data, albeit with reduced reconstruction quality.

In summary, the Monte Carlo study demonstrates that the addition of a second MAGIC telescope and the transition to stereoscopic observations significantly enhance angular and energy resolution, background rejection, and overall sensitivity from 60 GeV up to 20 TeV. These improvements open new opportunities for detailed spectral studies of faint sources, better localization of extended emission, and more efficient surveys of the very‑high‑energy sky. The authors conclude that forthcoming real‑data analyses will be essential to validate the simulated gains and to fine‑tune the analysis pipelines for optimal stereoscopic performance.