Extending the work of Czerny et al. (2004), we model the fractional variability amplitude due to distributions of hot spots co-orbiting on the accretion disk around a supermassive black hole. From defined radial distributions, our code samples random positions for the hot spots across the disk. The local spot emission is computed as reprocessed radiation coming from a compact primary source above the disk. The structure of the hot spot and the anisotropy of the re-emission are taken into account. We compute the fractional variability spectra expected from such spot ensembles and investigate dependencies on the parameters describing the radial spot distribution. We consider the fractional variability F_var with respect to the spectral mean and also the so-called point-to-point F_pp. Our method includes relativistic corrections for the curved space-time; the black hole angular momentum is a free parameter and subject to the fitting procedure. We confirm that the rms-variability spectra involve intrinsic randomness at a significant level when the number of flares appearing during the total observation time is too small. Furthermore, F_var is not always compatible with F_pp. For MCG -6-30-15, we can reproduce the short-timescale variability and model the suppressed variability in the energy range of the Kalpha line without any need to postulate reprocessing farther away from the center. An increasing rate of energy production by the flares toward the center of the disk, a fast rotation of the central black hole, and considerable suppression of the primary flare emission are required. The modeled line remains consistent with the measured equivalent width of the iron Kalpha line complex.