Are you an EPFL student looking for a semester project?
Work with us on data science and visualisation projects, and deploy your project as an app on top of GraphSearch.
Publication
Are you an EPFL student looking for a semester project?
Work with us on data science and visualisation projects, and deploy your project as an app on top of GraphSearch.
Many physically‐based models for climate change impact studies require sub‐daily temporal resolution of the forcing data to provide meaningful predictions. However, climate scenarios are typically available at daily time step, severely limiting the application of such physically‐based models. In this study, we propose an enhanced delta‐change method for downscaling climate change scenarios from daily to hourly resolution. The approach presented provides objective criteria for assessing the quality of the determined delta and downscaled time series, while also fixing issues of common quantile mapping methods used for spatial downscaling related to the decrease of correlation between different variables. However, this new approach has limitations in correctly representing statistically extreme events and changes in the frequency of discontinuous events such as precipitation. Smoothing of historical and future data is required prior to applying the delta‐change method, and the related parameters are found to have a subtle impact on the correctness of the representation of the seasonal means as well as the resulting (artificial) variability in the scenario data product. This new method is universal and can be applied with smoothing approaches apart from the harmonic fitting used in this work and in the past. In this study, the assessment suggested the use of seven harmonics for the smoothing of the input data as a best choice of this parameter for the data used. The method is applied to a Swiss climate change scenario data set, CH2018, and to a complement of this set to a Swiss alpine measurement network obtained by spatial transfer of CH2018, resulting in a set of 68 climate change scenarios at hourly resolution for 188 stations over Switzerland significantly expanding upon the spatial and temporal resolution of the CH2018 data set. All source code to perform such an analysis and the complete data product are provided open access.
, , ,