The influence of rooting depth on the simulated hydrological cycle of a GCM

Abstract

The hydrological cycle is one of the key issues in simulating the climate with an atmospheric general circulation model (GCM). This study shows that rooting depth and thus the field capacity (maximum soil waterholding capacity of the root zone) have a major influence on the simulated hydrological cycle of a GCM. We consider three different climate simulations of the GCM ECHAM4-T42: a control experiment using a dataset with spatially distributed field capacities, b) a simulation using optimized rooting depths in the tropics which are generally deeper than in the control experiment, c) a sensitivity simulation using a constant rooting depth of 60 cm in the tropics which is shallower than the corresponding average rooting depth of the control experiment. We evaluate the simulated hydrological cycles using the Hydrological Discharge model which simulates the lateral waterflows on a 0.5° grid using input fields from the GCM simulations. The model parameters depend on spatially distributed land surface characteristics. We compare the simulated discharges of a few large catchments to observed discharges as well as the simulated precipitation for these catchments to different climatologies. Here, we focus on the Sambesi catchment as a representative tropical catchment. For a specific catchment, differences in the hydrological cycle between the climate simulations may be directly based on differences in rooting depth due to a different evapotranspiration, or may be related to changes in the circulation patterns. The effects of rooting depth changes are discussed and conclusions for a future improvement of the representation of the hydrological cycle in atmospheric GCMs are drawn. The study reveals that the simulation with the deepest rooting depths is in general closest to observations.