In this paper we report on the application of statistical information obtained from 2D micrographs of catalyst layers (CL) of proton exchange membrane fuel cell (PEMFC), to derive stochastic replicas of their 3D pore networks at two scales. The main focus is on assessing the scaling strategy to determine effective transport coefficients and the analysis of the experimental results. The continuity equation for charge transport is solved directly on the 3D reconstructed CL to determine effective electrical conductivities at "internal-scale" and to simulate the electrical global performance at "macro-scale". The electrical performance is experimentally determined. The applied image processing method makes use of two immediate scales to improve the image resolution. Pore size distribution of the reconstructed scales is used to avoid the superposition of equal pore sizes. The relative porosities were determined by the statistical analysis of SEM micrographs and verified by mercury intrusion porosimetry.