In recent years, significant efforts have been made to upgrade physical processes in the ISBA‐CTRIP land surface system for use in fully coupled climate studies using the new CNRM‐CM6 climate model or in stand‐alone mode for global hydrological applications. Here we provide a thorough description of the new and improved processes implemented between the CMIP5 and CMIP6 versions of the model and evaluate the hydrology and thermal behavior of the model at the global scale. The soil scheme explicitly solves the one‐dimensional Fourier and Darcy laws throughout the soil, accounting for the dependency of hydraulic and thermal soil properties on soil organic carbon content. The snowpack is represented using a multilayer detailed internal‐process snow scheme. A two‐way dynamic flood scheme is added in which floodplains interact with the soil hydrology through reinfiltration of floodwater and with the overlying atmosphere through surface free‐water evaporation. Finally, groundwater processes are represented via a two‐dimensional diffusive unconfined aquifer scheme allowing upward capillarity rises into the superficial soil. This new system has been evaluated in off‐line mode using two different atmospheric forcings and against a large set of satellite estimates and in situ observations. While this study is not without weaknesses, its results show a real advance in modeling the physical aspects of the land surface with the new ISBA‐CTRIP version compared to the previous system. This increases our confidence that the model is able to represent the land surface physical processes accurately across the globe and in turn contribute to several important scientific and societal issues.

This paper presents a simple method for representing upward capillary fluxes from shallow groundwater into the unsaturated soil column of the large-scale hydrological models generally used at low resolution in global climate models. The groundwater scheme implemented in the Total Runoff Integrating Pathways river-routing model in a previous study is coupled with the Interaction between Soil Biosphere Atmosphere (ISBA) land surface model. In this coupling, the simulated water table depth acts as the lower boundary condition for the soil moisture diffusive equation. An original parameterization accounting for the subgrid topography inside each grid cell is proposed in order to compute this fully coupled soil lower boundary condition. The impact of this coupling on the simulated water budget is evaluated over France for the 1989-2009 period. Simulations are performed at high (1/12°) and low (0.5°) resolutions. Upward capillary fluxes induce a decrease in the simulated recharge from ISBA to the aquifers and contributes to an enhancement of the soil moisture memory. The simulated water table depths are then lowered, which induces a slight decrease in the simulated mean annual river discharges. These differences do not affect the comparison with observations. As a consequence, the simulated river discharges and water table heads compare still well with observations for the two soil bottom condition (free drain or fully coupled). It confirms the suitability of the coupling parameterization using subgrid spatial variability of topography. Compared to a free drain experiment, upward capillary fluxes at the bottom of the soil increase the mean annual evapotranspiration simulated over the aquifer domain by 3.12% and 1.54% at high and low resolutions, respectively. This increase can locally reach 50% and 30%, respectively.

Abstract. Groundwater is a non-negligible component of the global hydrological cycle, and its interaction with overlying unsaturated zones can influence water and energy fluxes between the land surface and the atmosphere. Despite its importance, groundwater is not yet represented in most climate models. In this paper, the simple groundwater scheme implemented in the Total Runoff Integrating Pathways (TRIP) river routing model is applied in off-line mode at global scale using a 0.5 • model resolution. The simulated river discharges are evaluated against a large dataset of about 3500 gauging stations compiled from the Global Data Runoff Center (GRDC) and other sources, while the terrestrial water storage (TWS) variations derived from the Gravity Recovery and Climate Experiment (GRACE) satellite mission help to evaluate the simulated TWS. The forcing fields (surface runoff and deep drainage) come from an independent simulation of the Interactions between Soil-Biosphere-Atmosphere (ISBA) land surface model covering the period from 1950 to 2008. Results show that groundwater improves the efficiency scores for about 70 % of the gauging stations and deteriorates them for 15 %. The simulated TWS are also in better agreement with the GRACE estimates. These results are mainly explained by the lag introduced by the lowfrequency variations of groundwater, which tend to shift and smooth the simulated river discharges and TWS. A sensitivity study on the global precipitation forcing used in ISBA to produce the forcing fields is also proposed. It shows that the groundwater scheme is not influenced by the uncertainties in precipitation data.