Impact de la variabilité non-mesurée des pré cipitations sur les débits en hydrologie urbaine: un cas d'étude dans le cadre multifractal

In this paper, the multifractal properties of a rainfall event, that occurred over the London area on February 9th 2009, are used to better understand and quantify the uncertainty associated with the rainfall variability not resolved by C-band radar, whose resolution is estimated at 1 km*1 km*5min, and how it is transferred to sewer discharge. The hydrological case study is the 900 ha urban catchment of Cranbrook (London). The multifractal properties of the rainfall field are in agreement with the simplest case of space-time multifractals that relies on a unique anisotropy exponent between space and time. This enables a multifractal downscaling of the rainfall field. First an ensemble of realistic downscaled rainfall fields is generated with the help of Universal Multifractals. Then the corresponding ensemble of hydrographs is simulated with the help of a semi-distributed urban model. It appears that the tail of probability distribution of the rainfall maximum (among the 100 samples generated) exhibit a power-law behaviour. The same statistical behaviour is observed for peak flows, but with different characteristic exponents.