Soot particle dynamics, including particle size distributions (PSDs) and related statistics, are of increasing practical significance due to evolving regulatory demands. The combination of a mass and number density preserving sectional model with a transported joint probability density function (JPDF) method ensures a full coupling of the joint scalar space, e.g. soot and gas phase reactions and radiative heat losses, within a method that can represent ignition/extinction phenomena as well as the slow (low Damköhler number) soot inception and oxidation chemistry in turbulent flames. This approach is here applied to the sooting non-premixed Sandia ethylene jet flame via a 78-dimensional joint-scalar space, including enthalpy, gas phase species and 62 soot sections. Soot nucleation is treated as a global step from acetylene to pyrene with the rate fitted using comparisons with full detailed chemistry. Soot surface growth is treated via a PAH analogy and soot oxidation is considered via O, OH and O2 using a Hertz–Knudsen approach. Comparisons with measured temperature, gas phase species and the mean soot volume fraction show good agreement while the introduction of zero soot diffusivity leads to substantially improved predictions of the RMS of the soot volume fraction. The calculated PSDs at the burner centreline show a transition from one to two-peaks along the axial direction with the mode of the second peak increasing from 14 to 32 nm. Scatter plots, joint statistics of soot parameters and temperature, and the chemical source terms across soot sections suggest that surface growth is dominant when PSDs are unimodal and that the competition of oxidation, coagulation/aggregation and surface growth leads to a PSD shape transition. It is also shown that local extinction events lead to the presence of soot in cool fuel lean mixtures.