Scale effect challenges in urban hydrology highlighted with a distributed hydrological model

Abstract

Abstract. Nowadays, hydrological models are extensively used in urban water management, future development scenario evaluation and research activities. A growing interest is devoted to the development of fully distributed and grid based models, following the increase of computation capabilities. The availability of high resolution GIS information is needed for such models implementation to understand flooding issues at very small scales. However, some complex issues about scaling effects still remain a serious issue in urban hydrology. The choice of an appropriate spatial resolution is a crucial problem, and the obtained model performance depends highly on the chosen implementation scale. In this paper we propose a two step investigation framework using scaling effects in urban hydrology. In the first step fractal tools are used to highlight the scale dependency observed within distributed data used to describe the catchment heterogeneity, both the structure of the sewer network and the distribution of impervious areas are analyzed. Then an intensive multi-scale modeling work is carried out to understand scaling effects on hydrological model performance. Investigations were conducted using a fully distributed and physically based model, Multi-Hydro, developed at Ecole des Ponts ParisTech (Multi-Hydro (2015)). The model was implemented at 17 spatial resolution ranging from 100 m to 5 m. Results coming out from this work demonstrate scale effect challenges in urban hydrology modeling. In fact, fractal concept highlights the scale dependency observed within distributed data used to implement hydrological models. Patterns of geophysical data change when we change the observation pixel size. The multi-scale modeling investigation performed with Multi-Hydro model at 17 spatial resolutions confirms scaling effect on hydrological model performance. Results were analyzed at three ranges of scales identified in the fractal analysis and confirmed in the modeling work. In the meantime, this work also discussed some issues remaining in urban hydrology modeling such as the availability of high quality data at higher resolutions and, model numerical instabilities as well as the computation time requirements. But still the principal findings of this paper allow replacing traditional methods of model calibration by innovative methods of model resolution alteration based on the spatial data variability and scaling of flows in urban hydrology.