Packed bed reactors with large particles and low aspect ratio are meaningful in engineering for pressure drop reduction and better heat removal. Experimental investigations have been carried out on the transient temperature variation and inhomogeneity in a packed bed reactor with an aspect ratio of 9.5 and filled with particles with an average diameter of 4.53 mm for methane-copper/copper oxide chemical looping combustion (CLC). The bed temperature evolution and spatial variation at axial and radial positions were measured and analyzed for oxidation reaction. The development of axial temperature front is found to be much slower than in the fine particle situation in the literature that is typical in laboratory reactors. The axial temperature profile is also apparently diffusion like, instead of being almost a discontinuously sudden change in fine particle reactors. The phenomena indicate that internal diffusion significantly influences heat transfer in the situation of the large particle size. Wall effects and radial temperature variation in this low aspect ratio reactor were analyzed as well; the reactor experiences strong non uniform radial temperature distribution, and channeling flow affects significantly the near wall temperature profile. A 2D unsteady pseudo homogeneous effective parameter model with radial and axial dispersion was developed to predict the transient heat transfer in the CLC process. A radially variable heat source term was added to the model to represent the effects of the non-uniform distribution of oxygen carrier material in the radial direction. The model predicts qualitatively reasonable 2D temperature distribution, especially the large radial temperature gradient which has been observed in the experiments. The predicted temperature front was a very abrupt change like a discontinuity face which is apparently different from the measured one in the experiments. Without including influences of the wall effects in the model, the used effective thermal conductivities were considered to be underestimated for conditions in the present study. (C) 2014 Elsevier Ltd. All rights reserved