The increase of urban areas within a catchment may affect its hydrological functioning in various ways. As expected, higher and anticipated peak flow enhanced by faster runoff from impervious surfaces and engineered drainage systems is often detected (Bhaduri et al., 2000; Burns et al., 2005). However, the decrease in base flow caused by reduced infiltration and sewer drainage was observed (Bhaduri et al., 2000; Joannis et al., 2002; Breil et al., 2010), but it was not necessarily clear in other cases (Burns et al., 2005; White and Greer, 2006). These inequalities could be partially explained by the difference in the spatial distribution of urban growth (Mejia and Moglen (2010a, b)), along with differences in storm water control devices, which may intensify or moderate the effects of increasing urban areas. The effects of increasing urban fraction being not easily predictable, modeling these changing processes remains an interesting challenge, particularly in lumped conceptual models. That is precisely what motivates this work, which consists in studying the impacts of the urban fraction on each of the four parameters of the lumped conceptual model GR4J. Two study sites for which the annual urbanized fraction values are available will be studied: the Ferson Creek at St. Charles and the Blackberry Creek at Yorkville, both in the USA. Two approaches are initially proposed: - Splitting the basin into a totally urbanized sub-catchment and a totally non-urbanized one. The rainfall is split between the two sub catchments according to the urbanized fraction, and the simulated discharges are combined by convolution. - Replacing the parameters that represent the reservoirs capacity and the unit hydrograph by functions depending on the urbanized fraction. We expect that the evaluation of these approaches may contribute to identify simple methodologies to deal with this particular aspect of non-stationary parameters that concerns urban sprawl in lumped models.