Accurate forecasting of volcanic particle (tephra) dispersal and fallout requires a reliable estimation of key Eruption Source Parameters (ESPs) such as the Mass Eruption Rate (Q_M). Q_M is usually estimated from the Top Plume Height (H_TP) using empirical and analytical models. For the first time, we combine estimates of H_TP and Q_M derived from the same sensor (radar) with mean wind velocity values (v_W) for lava-fountain fed tephra plumes associated with 32 paroxysms of Mt. Etna (Italy) to develop a new statistical model based on a Markov Chain Monte Carlo approach for model parameter estimation. This model is especially designed for application to radar data to quickly infer Q_M from observed H_TP and v_W, and estimate the associated uncertainty. It can be easily applied and adjusted to data retrieved by radars worldwide, improving our capacity to quickly estimate Q_M and related uncertainties required for the tephra dispersal hazard.