A recently proposed and successfully validated post-processing algorithm is extended to treat particle data obtained by a phase Doppler anemometer (PDA) in dense, fluctuating, low-speed, two-phase flow. In such a flow, Doppler signals may be noisy and split, since particles have long residence times in the measurement volume. A novel time average based on the measured burst lengths is proposed to estimate the cross-sectional area of the measurement volume. This time average accounts for the possible split of a noisy Doppler burst, and it provides a more general description of the measurement volume than the conventional ensemble average. The extended post-processing algorithm was tested on experimental time series obtained in a circulating fluidized bed and the algorithm was compared to conventional treatment of the particle data by a commercial PDA processor. The conventional processing strongly overestimates the vertical mass flux integrated over the cross-section. In contrast, time-mean and fluctuations of mass flux, particle volume concentration, and particle velocity are reliably estimated employing the proposed algorithm, with continuity check parameters in succeeding particle data, as well as with a particle velocity filter to estimate the size of the probe volume, applying a two-component PDA in a three-dimensional flow. (c) 2007 Elsevier Ltd. All rights reserved.