This work presents a revised model for microparticle detachment from a surface by turbulent air flow. The model accounts for the dependence of the effective Surface roughness at the microparticle/surface interface on the contact radius. This dependence is quantified by surface scanning using atomic force microscopy. A series of detachment experiments of glass microparticles, ranging from 30 to 110 mu m in diameter, from a glass substrate, was conducted to validate the revised model. The model and data were found to agree at the 95% confidence level. The model sensitivity to five different physical factors was analyzed using design-of-experiments methodology with a full-factorial-design layout. The uncertainty in the threshold-detachment free-stream velocity that arose from the inherent variability in the model inputs was quantified using Monte Carlo simulations. The uncertainty that occurs in the threshold-detachment free-stream velocity when there is no specific information on Surface roughness was estimated. (C) 2008 Elsevier Ltd. All rights reserved.