P C R - ­G L O B W B version 2.0: A High Resolution Integrated Global Hydrology and Water Resources Model

Abstract:

PCRaster GLOBal Water Balance is a grid-based global hydrological model developed at Utrecht University. It simulates soil moisture in vertically stacked soil layers, as well as exchange to the atmosphere and underlying groundwater reservoir. Fluxes are simulated under different land cover types by considering sub-grid variations in topography, vegetation phenology and soil properties. The model includes physically-based schemes for runoff generation and infiltration, resulting in direct runoff, interflow, groundwater recharge and baseflow, as well as channel routing.

We present the latest version of the model, PCR-GLOBWB 2.0, consolidating all new developments introduced since PCR-GLOWB 1.0 was first published (van Beek et al, 2011). The main new components are:

• An inclusion of water demand module and the progressive introduction of reservoirs and expansion of irrigation areas (Wada et al, 2014)
• An attribution of water use to ground- and surface water resources and the fate of return flow (de Graaf et al, 2014)
• A routing scheme accounting for variable extent of floodplains (Winsemius et al, 2013)

PCR-GLOBWB 2.0 now runs at a spatial resolution of 5 arc min (± 10 km) in comparison to the 30 arc min (50 km) resolution used in PCR-GLOWB 1.0. At the finer resolution and with the added components, PCR-GLOBWB 2.0 shows improvements over the previous version: observed discharges from 5142 GRDC stations can be approximated more closely and model efficiency improves, particularly for smaller catchment areas (ρ = 0.87); human impacts, altering the seasonal and inter-annual variation of terrestrial water storage, are well simulated and evident in the validation to GRACE data (ρ = 0.81). These improvements open up new possibilities to assess the state of global water resources.

Also, we show an outlook of model results at higher resolutions: 3 arc min (5 km) and 30 arc sec (1 km) for specific test-bed areas: California, Illinois and Rhine-Meuse. We discuss fundamental challenges in high resolution hydrological modeling and address various issues that range from computational challenges, lack of sufficient detail/fine information for parameterization - including atmospheric forcing - and emergent scaling problems at finer resolution (e.g. lateral exchange due to groundwater flow and water distribution network).