Abstract Accurate modelling of particulates build-up process is essential for designing effective stormwater management strategies. However, current modelling practice relies on the classical 'power model' which has limitations in accounting for the variability in the build-up process. This research study investigated the relationships between influential factors of the build-up process and coefficients in the power model. The outcomes showed that the coefficient, which determines the build-up rate, is predominantly influenced by land use factors (pervious area, road area, commercial area and residential area), such that land use factors exerted 23 times more influence than the site characteristics (distance to pervious area and road surface texture depth). The coefficient, which determines how quickly build-up reaches equilibrium, was found to be equally influenced by anthropogenic activities (sweeping frequency and traffic volume) and site characteristics. Further, site characteristics were found to play a major role in generating build-up process variability with three times more influence than that of anthropogenic activities. It was found that the power model satisfactorily replicates the build-up of particles <74 μm. For the build-up of particles >74 μm, a new coefficient, namely, 'coefficient of variability' was introduced in order to improve the prediction performance (up to 17% compared to original power model). The study outcomes provide a deeper understanding into particulates build-up modelling, and can contribute to the formulation of effective stormwater treatment strategies. Graphical abstract Unlabelled Image Highlights • Classical power model satisfactorily replicates build-up of fine particles. • A new 'coefficient of variability' better replicates build-up of coarser particles. • Coefficients of build-up rate and build-up curve were defined. • Site characteristics largely influence the variability due to particle behaviour. [ABSTRACT FROM AUTHOR]