ORCHIDEE-ROUTING: revising the river routing scheme using a high-resolution hydrological database

Trung Nguyen-Quang,Ana Schneider,Lluís Fita,Xudong Zhou,Jan Polcher,Agnès Ducharne,Thomas Arsouze

doi:10.5194/gmd-11-4965-2018

Abstract

Abstract. The river routing scheme (RRS) in the Organising Carbon and Hydrology in Dynamic Ecosystems (ORCHIDEE) land surface model is a valuable tool for closing the water cycle in a coupled environment and for validating the model performance. This study presents a revision of the RRS of the ORCHIDEE model that aims to benefit from the high-resolution topography provided by the Hydrological data and maps based on SHuttle Elevation Derivatives at multiple Scales (HydroSHEDS), which is processed to a resolution of approximately 1 km. Adapting a new algorithm to construct river networks, the new RRS in ORCHIDEE allows for the preservation of as much of the hydrological information from HydroSHEDS as the user requires. The evaluation focuses on 12 rivers of contrasting size and climate which contribute freshwater to the Mediterranean Sea. First, the numerical aspect of the new RRS is investigated, in order to identify the practical configuration offering the best trade-off between computational cost and simulation quality for ensuing validations. Second, the performance of the new scheme is evaluated against observations at both monthly and daily timescales. The new RRS satisfactorily captures the seasonal variability of river discharge, although important biases stem from the water budget simulated by the ORCHIDEE model. The results highlight that realistic streamflow simulations require accurate precipitation forcing data and a precise river catchment description over a wide range of scales, as permitted by the new RRS. Detailed analyses at the daily timescale show the promising performance of this high-resolution RRS with respect to replicating river flow variation at various frequencies. Furthermore, this RRS may also eventually be well adapted for further developments in the ORCHIDEE land surface model to assess anthropogenic impacts on river processes (e.g. damming for irrigation operation).

Highlights

Large-scale river routing is a valuable tool for validating the performance of land surface models (LSMs)
This study presents an attempt to improve the river routing scheme in the ORCHIDEE LSM, in order to benefit from the accurate hydrography provided by the HydroSHEDS database at a 1 km resolution
This high-resolution information is aggregated in hydrological transfer units which are constructed inside each ORCHIDEE grid box

Summary

Introduction

Large-scale river routing is a valuable tool for validating the performance of land surface models (LSMs). The importance of river routing schemes (RRSs) in LSMs to properly estimate the land water storage variation has been demonstrated by Ngo-Duc et al (2007a). The absence of flow routing in LSMs confines the evaluation of run-off simulation to medium sized catchments (Beck et al, 2016) or an individual basin with a crude estimated water residence time (Balsamo et al, 2009). This limits the benefit of river discharge measurements which provide a unique accurate signal of the continental water cycle (Fekete et al, 2012). The representation of the lateral transport of water is recognized as an important topic in LSM development as it contributes directly to closing the water cycle in Earth system models (e.g. Arora et al, 1999; Ngo-Duc et al, 2007b; Gong et al, 2011; David et al, 2011) or in a fully coupled atmosphere–land–ocean model

Objectives

Results

Discussion

Conclusion