A new methodology to propagate uncertainties from regional climate models to urban impact model : feasibility study

The demand from the city planners and stakeholders concerning climate change impact on cities is increasing. Due to the global warming, cities, which are already more vulnerable, will experience an increasing number of extreme events such as heat waves. However, the information about long-term climate projections are extracted from global or regional climate models that are not conducted at city-scale resolutions. Additional simulations are therefore required. The computational cost of running a high-resolution regional climate model (RCM) coupled with a land surface model (LSM) to get the full atmospheric feedback from cities is extremely high. Moreover, an ensemble of simulations is required in order to estimate the uncertainties in part related to urban projections. Using a LSM in standalone mode is a cheaper solution. However, to downscale climate simulations from a RCM without urban parameterisation to a LSM, an extra step is required in order to present accurate results to compensate for the lack of atmosphere-land feedback. A simple and relatively fast statistical–dynamical downscaling approach is developed to fulfil this step and correct the LSM forcing. In this study, a description and a validation of the methodology is presented. The RCM used is ALARO, and the LSM is SURFEX. They are run over the Brussels Capital Region and forced with ERA-Interim reanalysis during the summers of 1981–2010. The validation is performed for the urban heat island