Keywords
Energy budget | Land surface modeling | Empirical Multistructure Framework (EMF)
No hydrological model in existence today can convincingly claim that it is systematically superior to the others, for all types of applications and under all conditions. Our contribution to model development is structured twofold. 1) For climate change studies, it is deemed important to account for the energy budget and not only for the water budget, since evaporation processes are the only hydrometeorological flux influencing both. This may be achieved through the use of a land surface model or of a surface flux partitioning model based on maximum entropy production. 2) For simulation or forecast, the combinatory usage of many simple models turns out to be very competitive. The Empirical Multistructure Framework (EMF), designed by our team, allows exploring a very large number of such simple models.
Selecting an appropriate objective function allows optimizing the model through calibration for some specific streamflow quantiles
Simulations of the soil water content from two land surface models are compared to observations
Performance (MCRPS) of 1500 models, generated by the EMF, are compared before and after EnKF data assimilation. Results are presented for individual and global (EMF) calibrations