The need for hydrological modeling in ungauged basins has always been a major challenge. Efforts have been made in recent years to address it via the implementation of ex-situ technologies. An attractive option that is becoming increasingly popular are reanalysis products, which provide modeled data for atmospheric variables, based on observations assimilated from various sources, such as satellites and radar stations. The present study is focusing on the evaluation of two reanalysis precipitation and temperature datasets, namely ERA5 and COSMO-REA6, over a selection of drainage basins located in Slovenia. The two products are available at the hourly temporal scale, and at the time of writing, their quality has not been evaluated in rainfall-runoff applications over European borders. To measure their efficiency, the GR4H and GR4H Cema Neige lumped conceptual rainfall-runoff models were used. Different reanalysis configurations were tested by using the initial model parameters, calibrated using observed data. In addition, it was investigated whether the parameters ingrained within the model are able to make up for low rainfall and temperature data quality, by recalibrating them on every reanalysis configuration. Initial model runs proved to be adequate only for a few catchments out of the total selection. However, the recalibration process was effective in increasing hydrological performance, increasing the total number of watersheds that were modeled accurately. The study concluded that the two products are partially in agreement with observations, suggesting their use as a proxy when observed data is scarce.
|
