The central engine causing the production of jets in radio sources may work intermittently, accelerating shells of plasma with different mass, energy and velocity. Faster but later shells can then catch up slower earlier ones. In the resulting collisions shocks develop, converting some of the ordered bulk kinetic energy into magnetic field and random energy of the electrons which then radiate. We propose that this internal shock scenario, which is the scenario generally thought to explain the observed gamma-ray burst radiation, can work also for radio sources in general, and for blazar in particular. We investigate in detail this idea, simulating the birth, propagation and collision of shells, calculating the spectrum produced in each collision, and summing the locally produced spectra from those regions of the jet which are simultaneously active in the observer's frame. We can thus construct snapshots of the overall spectral energy distribution as well as time dependent spectra and light curves. This allows us to characterize the predicted variability at any frequency, study correlations among the emission at different frequencies, specify the contribution of each region of the jet to the total emission, find correlations between flares at high energies and the birth of superluminal radio knots and/or radio flares. The model has been applied to qualitatively reproduce the observed properties of 3C 279. Global agreement in terms of both spectra and temporal evolution is found. In a forthcoming work, we explore the constraints which this scenario sets on the initial conditions of the plasma injected in the jet and the shock dissipation for different classes of blazars.Comment: 12 pages, 10 postscript figures, submitted to MNRA

We present an analysis of the Internal Shock Model of GRBs, where gamma-rays are produced by internal shocks within a relativistic wind. We show that observed GRB characteristics impose stringent constraints on wind and source parameters. We find that a significant fraction, of order 20%, of the wind kinetic energy can be converted to radiation, provided the distribution of Lorentz factors within the wind has a large variance and provided the minimum Lorentz factor is > Γ ± ≈ 10 2.5 L 2/9 52 , where L = 10 52 L 52 erg s −1 is the wind luminosity. For a high, > 10%, efficiency wind, spectral energy breaks in the 0.1 to 1 MeV range are obtained for sources with dynamical time R/c ∼ < 1 ms, suggesting a possible explanation for the observed clustering of spectral break energies in this range. The lower limit Γ ± to wind Lorenz factor and the upper limit ≈ 1(R/10 7 cm) −5/6 MeV to observed break energies are set by Thomson optical depth due to e ± pairs produced by synchrotron photons. Natural consequences of the model are absence of bursts with peak emission energy significantly exceeding 1 MeV, and existence of low luminosity bursts with low, 1 keV to 10 keV, break energies.Subject headings: gamma-rays: bursts -methods: numerical -radiation mechanisms: non-thermal Both the spectrum and temporal dependence of afterglow emission are consistent with synchrotron emission of electrons accelerated to high energy at the shock wave driven by the

We simulate Gamma-Ray Bursts arising from internal shocks in relativistic winds, calculate their power density spectrum (PDS), and identify the factors to which the PDS is most sensitive: the wind ejection features, which determine the wind dynamics and its optical thickness, and the energy release parameters, which give the pulse 50-300 keV radiative efficiency. For certain combinations of ejection features and wind parameters the resulting PDS exhibits the features observed in real bursts. We found that the upper limit on the efficiency of conversion of wind kinetic energy into 50-300 keV photons is $\sim$ 1%. Winds with a modulated Lorentz factor distribution of the ejecta yield PDSs in accord with current observations and have efficiencies closer to $10^{-3}$, while winds with a random, uniform Lorentz factor ejection must be optically thick to the short duration pulses to produce correct PDSs, and have an overall efficiency around $10^{-4}$.Comment: 6 pages, 4 figures, Latex, submitted to The Astrophysical Journal (05/04/99