Premium
Global‐scale river routing—an efficient time‐delay algorithm based on HydroSHEDS high‐resolution hydrography
Author(s) -
Gong L.,
Halldin S.,
Xu C.Y.
Publication year - 2011
Publication title -
hydrological processes
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.222
H-Index - 161
eISSN - 1099-1085
pISSN - 0885-6087
DOI - 10.1002/hyp.7795
Subject(s) - routing (electronic design automation) , computer science , environmental science , hydrography , algorithm , streamflow , surface runoff , remote sensing , drainage basin , geology , geography , computer network , ecology , oceanography , cartography , biology
Abstract Coupling of global hydrologic and atmospheric models is difficult because of the highly nonlinear hydrological processes to be integrated at large scales. Aggregation of high‐resolution data into low‐resolution spatial distribution functions is one way to preserve information and account for the nonlinearity. We used HydroSHEDS, presently the most highly resolved (3″) global hydrography available, to provide accurate control on global river routing through a computationally efficient algorithm. The high resolution of HydroSHEDS allowed discrimination of river‐channel pixels, and time‐delay distributions were calculated for all such pixels. The distributions were aggregated into network‐response functions (NRFs) for each low‐resolution cell using an algorithm originally developed for the 1‐km‐resolution HYDRO1k hydrography. The large size of HydroSHEDS required a modification in algorithm to maintain computational efficiency. The new algorithm was tested with a high‐quality local and a more uncertain global weather dataset to identify whether improved routing would provide a gain when weather data quality was limiting. The routing was coupled to the WASMOD‐M runoff‐generation model to evaluate discharge from the Dongjiang River and the Willamette River basins. The HydroSHEDS‐based routing, compared with the HYDRO1k‐based routing, provided a small gain in model efficiency, for local and global weather data and for both test basins. The HydroSHEDS‐based routing, contrary to the HYDRO1k‐based routing, provided physically realistic wave velocities. The most stable runoff‐generation parameter values were achieved when HydroSHEDS was used to derive the NRFs. Routing was computed in two steps: first, a preparatory calculation which was a one‐time effort and second, the routing during each simulation. The computational efficiency was four to five orders of magnitude better for the simulation step than that for the preparatory step. Copyright © 2010 John Wiley & Sons, Ltd.