Evaluation of simulated discharge in global hydrological models from ISIMIP3a

Climate change is expected to have a strong impact on the hydrological cycle, resulting in a higher prevalence of hydrological extremes, including river floods and hydrological droughts. Global hydrological models (GHMs) are widely used to assess changes in river flow. We evaluated nine GHMs that contributed simulations to ISIMIP3a (CLASSIC, CWatM, H08, HydroPy, JULES-W2, MIROC-INTEG-LAND, ORCHIDEE-MICT, WaterGAP2-2e, and WEB-DHM-SG). Models used the same observation-based atmospheric climate forcing (GSWP3-W5E5, Lange et al. 2022) provided by ISIMIP and direct human forcings (e.g., land use, river routing, water abstraction). To evaluate model performance, we used observed discharge from the Global Runoff Data Centre (GRDC). Model performance was evaluated for 644 stations, which have been shown to be suitable for evaluating river discharge data from ISIMIP GHMs (Schmied and Schiebener 2022). River routing schemes differ between the modelling groups, and it has been shown that discharge modelled using the river routing model CaMa-Flood can outperform the models’ internal routing scheme (Zhao et al. 2017). In flood modelling, CaMa-Flood is often used in combination with the output from GHMs. For this study, we compared the observed discharge with the discharge simulated by the model’s internal routing scheme, and the discharge simulated by CaMa-Flood using the runoff simulated by each GHM. We used daily and maximum annual discharge simulated on a 0.5°x0.5° grid and for the years 1901-2019. The figures cover the time-period for which observational data was available.

---

Location of the 644 stations that were used for this evaluation study

---

Link to isimip data

---

FIG 1