Global assessment of hydrological components using a seamless multiscale modelling system

Our study provides a global assessment of water balance components accounting for the uncertainty in globally available precipitation products. This assessment is carried out consistently using a multiscale modelling framework established over more than 6000 GRDC river basins at various spatial resolutions (daily time step, period 1990-2019). The framework is based on the mesoscale Hydrologic Model (mHM; [1,2,3]) equipped with the multiscale parameter regionalization (MPR) scheme. All basins share the same parameterization and are driven with four different state-of-the-art meteorological products: ERA5 reanalysis [4], MSWEP [5], and deterministic EM-Earth v2 [6]. Additionally, the hydrological simulations are benchmarked against E-OBS [7] over Europe and against locally interpolated 1km gridded rain gauge dataset over Germany.

Our results show that EM-Earth clearly exhibits the best streamflow performance across North and South America and Asia with respect to the other products. The MSWEP is the best product in Africa, where the overall model’s performance is rather poor. In Australia, MSWEP and EM-Earth have comparable skills. In Europe, the differences get narrower, although slightly better performance is seen for EM-Earth, with a median performance of 0.5 of daily KGE (N basins=2000), mainly due to correcting the under-catch error of rain gauges which is not considered, e.g. in E-OBS. Furthermore, when we zoom into a subset consisting of medium-sized 200 German basins, the in-house high-resolution meteorologic product clearly overperformed all global products, mainly due to better captured temporal correlation and smaller biases. On the other hand, ERA5 leads to very strong positive biases over German basins using standard parameterization. Finally, our study contributes to discussions on objective quantification of the optimal spatial resolution of hydrological studies.

 

[1] https://doi.org/10.5281/zenodo.5119952

[2] https://agupubs.onlinelibrary.wiley.com/doi/full/10.1029/2008WR007327 

[3] https://agupubs.onlinelibrary.wiley.com/doi/full/10.1029/2012WR012195

[4] https://rmets.onlinelibrary.wiley.com/doi/10.1002/qj.3803

[5] https://hess.copernicus.org/articles/21/589/2017/hess-21-589-2017.html

[6] https://journals.ametsoc.org/view/journals/bams/103/4/BAMS-D-21-0106.1.xml

[7] https://agupubs.onlinelibrary.wiley.com/doi/full/10.1029/2017JD028200