Sediment reworking rates by the surface deposit-feeding bivalve Abra ovata were assessed in thin aquaria using an automated image analysis procedure for luminophore tracer particles. Experiments were carried in winter (10 °C) and summer (20 °C), and three food treatments were tested: no food addition (CF), low food addition (LF) and high food addition (HF). The rate of sediment reworking was characterized in four different ways: (1) the maximum penetration depth of luminophores (MPD), (2) the proportion of reworked sediment surface (PRS) from which individual rates of surface area reworking (IRSAR) were derived, (3) the fitting of the standard biodiffusion model resulting in a biodiffusion coefficient Db, and (4) the application of a new non-local bioturbation model that produced an alternative mixing intensity DbNL. In winter, sediment reworking was low and was not affected by food availability. In contrast, during summer, reworking activity was very high and significantly affected by food availability. This suggests that temperature and not food availability controls sediment reworking during wintertime. Although the biodiffusive and non-local models produced similar values for mixing intensities after 48 h, the non-local model gave markedly better fits during the initial stage of the experiment. This agrees nicely with theoretical predictions: over short-time scales the non-local model should provide a more accurate description of bioturbation, but as the number of bioturbation events increases, the non-local model should converge to the biodiffusion one. Yet, an additional advantage of the non-local model is that it allows constraining two crucial parameters characterizing A. ovata bioturbation: the average distance over which particles are displaced (2.1 mm) and the waiting time between two bioturbation events (5.39 h). Accordingly, reworking is characterized by highly frequent and small-scale particle displacement, which makes that A. ovata can be classified as a true biodiffuser.