Long term storage of pear fruit requires low temperature and conventionally uses controlled atmosphere (CA) conditions to reduce respiration and consequent quality loss. Sub-optimal storage conditions may lead to physiological disorders and loss of product. Stochastic variability of the properties of fruit introduces uncertainty in storage design and operations and could result in severe quality loss. Taking such variability into account in simulation models for virtual engineering will allow to assess the uncertainty of the process and determine confidence limits for the operation. Gas exchange in pear fruit during controlled atmosphere storage was studied using a continuum diffusion respiration model, taking into account stochastic variation of the 3D morphology, the diffusivity of oxygen and carbon dioxide and the maximal respiration rate. Different geometries were generated using a statistical shape generation algorithm for 3D morphology, that was automatically incorporated into the gas exchange model. Similarly, tissue diffusivity was computed using a 3D tissue microstructure database. Simulation results showed that internal O-2 and CO2 gas profiles in fruit were highly affected by variation of diffusivities, maximal respiration rate and the 3D morphology of fruit. The model was further used to evaluate incidence to fermentation at different reduced 02 levels of storage condition. The risk of fermentation inside the fruit predicted by the gas exchange model rapidly increased in response to decreasing external 02 levels. The virtual simulation tool confirms that picking time and fruit size are important criteria for proper control of CA storage. While applied here to pear fruit, it can easily be extended to other commodities. (C) 2015 Elsevier Ltd. All rights reserved.