Previous studies in the literature have shown experimentally that the critical current density in bulk MgB2 depends on the position at which it is measured within the sample, and that the trapped magnetic field saturates above large values of sample diameter. To better understand this behaviour, we have carried out a detailed study of the trapped magnetic field and local critical current distribution in bulk MgB2 using numerical modelling with H-formulation. The properties of bulk MgB2 samples with radiusindependent (i.e. uniform) and radius-dependent critical current were modelled, based on experimentally-determined parameters obtained from the literature. Radius-independent samples took properties originally derived from either the central or edge region of the original samples. In our study it was determined that the peak trapped magnetic field value of the sample with the central region properties was higher by 22.4% and 9.20% than the samples with edge region properties and that with radius dependent parameter, respectively. Additionally, to understand why the trapped magnetic field saturates as the diameter of sample increases, the trapped magnetic field was modelled using the radius dependent critical current density J(c) (B, r) as input data rather than using a constant bulk critical current density. The peak trapped magnetic field values increase by 8.6% and 6.2% respectively for the central region radius-independent and radius-dependent property bulk MgB2 samples as the bulk diameter increases from 25 mm to 35 mm. This tendency to saturation in the peak trapped field indicates that increasing the bulk diameter alone does not have a significant effect on the value of the trapped magnetic field, unless the bulk superconducting current density is improved uniformly throughout the MgB2 bulk. These results enable us to understand how we can experimentally enhance the trapped magnetic field in bulk MgB2 by producing samples with uniform high critical current density distribution using fabrication methods such as graded doping. (C) 2016 Elsevier B. V. All rights reserved.