Assessing the impact of the development of an urban district on shallow groundwater using the integrated urban hydrological model URBS

The urbanisation leads to modifications in the water budget, not only at the surface but in groundwater as well. Few urban modelling studies deal with this topic, due to the lack of appropriate models. The URBS (Urban Runoff Branching Structure) model has been developed since several decades to simulate water transfers at the scale of an urban district. An integrated modelling approach is deliberately adopted to account for the numerous elements that influence urban hydrology: the spatial distribution of the sealed surfaces, interactions between the urban soil and water networks or underground, sustainable drainage systems…. In URBS, the spatial discretization of a catchment is based on Urban Hydrologic Elements (UHE) constituted by cadastral parcels and the adjacent streets, connected to the drainage network. URBS is able to perform continuous and long-period simulations (typically several years) of water fluxes in urban districts for small time-steps (typically few-minutes), with rainfall and potential evapotranspiration as input data.

The URBS model is adopted to study the hydrological impact of the Moulon district layout, a 200 ha development operation of the Paris-Saclay Cluster (currently underway). The project should result in an increase of sealed surfaces from 14% to 35% and a densification of underground constructions such as networks and basements. A shallow unconfined aquifer extends on the whole area. The fluctuations of ground-water levels have been monitored at an hourly time-step with 8 piezometers since 2012. Water-table levels exhibit significant variations, with near-saturation levels during winter and several meters depths during summer, although the piezometers do not all exhibit the same dynamics.

A calibration of the URBS model is first conducted for a 2-year period using only piezometric data and no flowrate data. The calibration is solely performed for the parameters influencing the soil compartment: soil permeability and parameters of the sewer infiltration process. Model performances are rather satisfactory with good representation of the observed levels for several piezometers, despite some difficulties for two piezometers exhibiting atypical variations. Once the URBS model is calibrated for the initial situation, simulations are conducted for the project layout (accounting for land-use modification and underground constructions) so as to evaluate the hydrological impacts of the development. Simulation results suggest that an increase of water table levels might be expected after the development of the district (this somehow surprising result may partly originate from the decrease of evapotranspiration fluxes associated with the increased of sealed surfaces).

The analysis of these first simulations also suggests that large uncertainties might be expected regarding the water levels computed by URBS. A simplified uncertainty analysis (based on Monte-Carlo simulations) is thus conducted to evaluate and distinguish uncertainties associated with model parameters and the total uncertainties in model outputs. While the results clearly evidence the importance of total uncertainties (although the uncertainties due to the model parameters remain low), they also confirm that groundwater depths could be reduced by the construction of the Moulon district.