I Congreso de Postgrado fcfm: ingeniería, ciencias e innovación

185 Santiago, 10 al 12 de agosto, 2022 TO WHAT EXTENT DOES RIVER ROUTING MATTER IN HYDROLOGICAL MODELING? Nicolás Cortes-Salazar 1* , Nicolas Vasquez 1 , Naoki Mizukami 2 , Pablo A. Mendoza 1,3 and Ximena Vargas 1 ¹Department of Civil Engineering, Universidad de Chile, Santiago, Chile. ²National Center for Atmospheric Research (NCAR), Boulder, Colorado, USA. ³Advanced Mining Technology Center (AMTC), Universidad de Chile, Santiago, Chile. *Email: nicolas.cortes.s@ug.uchile.cl ABSTRACT Spatially-distributed hydrology and land surface models are typically applied in combination with river routing schemes that convert instantaneous runoff into discharge. Nevertheless, the development of such schemes has been somehow disconnected from hydrologic model calibration research [1], although both seek to achieve more realistic streamf low simulations. In this paper, we seek to bridge this gap to understand the extent to which the configuration of routing schemes affects hydrologic model calibration results in water resources applications. To this end, we configure the Variable Infiltration Capacity (VIC) [2], coupled with the mizuRoute routing model [3] in the Cautin river basin (2770 km2), Chile. We use the Latin Hypercube Sampling (LHS) method to generate 3500 different VIC model parameters sets, for which basin-averaged runoff estimates are obtained directly (no rout- ing case), and subsequently compared against outputs from four routing schemes (Unit Hydrograph, Lagrang- ian Kinematic Wave, Muskingum-Cunge and Diffusive Wave) applied with five different routing time steps (1, 3, 6, 12 and 24 hours). The results show that incorporating routing schemes may alter streamf low simulations at sub-daily, daily and even monthly time scales. The maximum Kling-Gupta Efficiency (KGE) [4] obtained for daily discharge increases from 0.73 (no routing) to 0.82 (for the best scheme), and such improvements do not depend on the routing time step. Moreover, the optimal parameter sets may differ depending on the routing scheme con- figuration, affecting the basef low contribution to total runoff. Including routing models decreases streamf low values in frequency curves and the segment with high discharge values in the f low duration curve (compared to the case without routing). More generally, the results presented here highlight the impact that river routing implementations may have for water resources applications that involve hydrologic models and, in particular, parameter calibration. ACKNOWLEDGMENTS This work was partially funded by the project: “Fondecyt de iniciación en investigación 11200142”. REFERENCES [1] A. J. Newman, A. G. Stone, M. Saharia, K. D. Holman, N. Addor, and M. P. Clark. Hydrol. Earth Syst. Sci., vol. 25, 5603 (2021) [2] X. Liang, D. P. Lettenmaier, E. F. Wood, and S. J. Burges. J. Geophys. Res., vol. 99 , 14415 (1994) [3] N. Mizukami et al., Geosci. Model Dev., vol. 9 , 2223 (2016) [4] H. V. Gupta, H. Kling, K. K. Yilmaz, and G. F. Martinez. J. Hydrol., vol. 377 , 80 (2009) 16 R E CU R SOS H I D R I COS