Large-amplitude late-time radio variability in GRB 151027B

Context. Deriving physical parameters from gamma-ray burst (GRB) afterglow observations remains a challenge, even 20 years after the discovery of afterglows. The main reason for the lack of progress is that the peak of the synchrotron emission is in the sub-mm range, thus requiring radio observations in conjunction with X-ray/optical/near-infrared data in order to measure the corresponding spectral slopes and consequently remove the ambiguity with respect to slow vs. fast cooling and the ordering of the characteristic frequencies. Aims. We have embarked on a multifrequency, multi-epoch observing campaign to obtain sufficient data for a given GRB that allows us to test the simplest version of the fireball afterglow model. Methods. We observed GRB 151027B, the 1000th Swift-detected GRB, with GROND in the optical-near-IR, ALMA in the sub-millimeter, ATCA in the radio band; we combined this with public Swift/XRT X-ray data. Results. While some observations at crucial times only return upper limits or surprising features, the fireball model is narrowly constrained by our data set, and allows us to draw a consistent picture with a fully determined parameter set. Surprisingly, we find rapid, large-amplitude flux density variations in the radio band which are extreme not only for GRBs, but generally for any radio source. We interpret them as scintillation effects, though their extreme nature requires the scattering screen to be at a much smaller distance than usually assumed, multiple screens, or a combination of the two. Conclusions. The data are consistent with the simplest fireball scenario for a blast wave moving into a constant-density medium, and slow-cooling electrons. All fireball parameters are constrained at or better than a factor of 2, except for the density and the fraction of the energy in the magnetic field which has a factor of 10 uncertainty in both directions.