Decoding FRB energetics and frequency features hidden by observational incompleteness

Fast radio bursts (FRBs) are millisecond-duration radio flashes of extragalactic origin, with magnetars implicated as viable central engines. Yet their triggering and radiation mechanisms remain unknown. Radio telescopes inevitably record bursts incompletely, as limited sensitivity and finite bandwidth lead to observational truncation. Here we establish a general analytical framework to reconstruct intrinsic population-level frequency characteristics and energetic parameters directly from observationally truncated FRB data. Applying this method to 2,223 bursts of FRB 20121102A observed by three different telescopes, we show that the narrow spectra of repeating FRBs are predominantly an observational selection effect. Only intrinsically high-energy bursts are genuinely narrowband. We further quantify, for the first time, the number and energy of completely undetected bursts, and reveal intrinsic long-term frequency evolution of the source. Our methodology transforms incomplete archival observations into physically meaningful probes, bridging instrumental readouts and intrinsic FRB physics.

💡 Research Summary

The authors address a fundamental problem in fast radio burst (FRB) studies: the incompleteness of observed bursts caused by limited telescope sensitivity and finite observing bandwidths. Using a large sample of 2,223 bursts from the repeating source FRB 20121102A, recorded with Arecibo, the Green Bank Telescope (GBT), and FAST, they develop a general analytical framework that reconstructs the intrinsic spectral and energetic properties of the population directly from truncated data.

The core of the method assumes that each burst’s spectral energy distribution (SED) is well described by a Gaussian function. By relating the observed fluence integrated over the instrument’s band (F_eqv,ν,obs) and the observed bandwidth (δν_obs) to the intrinsic peak frequency (ν_p) and intrinsic full‑width‑half‑maximum (σ_ν), they derive two key equations. Equation (1) connects the observed fluence to the cumulative Gaussian function, while Equation (2) gives the total fluence in terms of σ_ν and the fluence threshold. Solving these equations with an adaptive gradient‑descent algorithm yields ν_p and σ_ν for each burst without performing a full spectral fit, which would be unreliable for weak events or for bursts whose peaks lie outside the band.

The sample is divided into “band‑unlimited” bursts (those whose spectra are fully contained within the observing band) and “band‑limited” bursts (those truncated by the band edges). For the 113 band‑unlimited bursts from Arecibo, the authors recover intrinsic ν_p and σ_ν, then explore correlations with observed quantities such as fluence, waiting time, and box‑car equivalent pulse width. They find a near‑linear relation F_eqv,ν,obs ≈ 0.0044 F^0.94, confirming that using the full instrumental bandwidth as a proxy for burst bandwidth is justified for these events. Moreover, they uncover a scaling F ∝ W_eqv,obs^2, implying that higher‑energy bursts tend to have longer durations.

For the majority (≈75 %) of bursts that are band‑limited, the authors model the intrinsic lower and upper spectral bounds (ν_ℓ, ν_h) as a bivariate Gaussian distribution. Bayesian inference, combined with the observed truncation imposed by the instrumental band (f_ℓ, f_h), allows them to reconstruct the underlying (ν_ℓ, ν_h) distribution and, by linear transformation, the intrinsic (ν_p, δν) distribution. The reconstructed intrinsic bandwidth distribution peaks at roughly twice the observed peak, demonstrating that the narrow spectra commonly reported for repeating FRBs are largely a selection effect. Only intrinsically high‑energy bursts retain genuinely narrow bandwidths; low‑energy bursts appear narrow because their spectral wings fall below the fluence threshold.

Quantitatively, the model predicts that about 1.64 % of all bursts from FRB 20121102A are completely undetected because their entire spectra lie outside the observing band. Conditional on detection, there is a 73.9 % probability that a burst is affected by the operating‑band cutoff, with the dominant scenario being ν_ℓ < f_ℓ < f_h < ν_h.

The authors also fit shifted log‑normal distributions to the intrinsic and observed energy populations using Markov Chain Monte Carlo (MCMC) simulations. They find that sensitivity truncation removes roughly 5 % of the total burst energy in the band‑unlimited sample, but it dramatically inflates the low‑energy tail of the observed distribution. This correction is essential for accurate estimates of the FRB energy function and event rate.

Finally, by tracking the reconstructed ν_p over the multi‑year observing span, the study reveals a gradual long‑term drift in the peak frequency, suggesting intrinsic spectral evolution of the source. This behavior is consistent with magnetar models where the emission region or magnetic field configuration evolves over time.

In summary, the paper provides a robust, telescope‑agnostic methodology to invert observational truncation effects and recover the true energetic and spectral characteristics of FRBs. It demonstrates that the apparent narrowness of repeating FRB spectra is mostly an artifact of limited sensitivity and bandwidth, that only the most energetic bursts are intrinsically narrowband, and that a non‑negligible fraction of bursts remain completely unseen. These insights refine our understanding of FRB energetics, occurrence rates, and the physical mechanisms powering them, and they lay the groundwork for applying similar corrections to other FRB samples, including those from upcoming wide‑field facilities.