The multiscale routing model mRM v1.0: simple river routing at resolutions from 1 to 50 km

Abstract

Abstract. Routing streamflow through a river network is a fundamental\nrequirement to verify lateral water fluxes simulated by hydrologic\nand land surface models. River routing is performed at diverse\nresolutions ranging from few kilometres to 1 ∘ . The presented\nmultiscale routing model mRM calculates streamflow at diverse\nspatial and temporal resolutions. mRM solves the kinematic wave\nequation using a finite difference scheme. An adaptive time stepping\nscheme fulfilling a numerical stability criterion is introduced in\nthis study and compared against the original parameterisation of mRM\nthat has been developed within the mesoscale hydrologic model\n(mHM). mRM requires a high-resolution river network, which is\nupscaled internally to the desired spatial resolution. The user can\nchange the spatial resolution by simply changing a single number in\nthe configuration file without any further adjustments of the input\ndata. The performance of mRM is investigated on two datasets: a\nhigh-resolution German dataset and a slightly lower resolved\nEuropean dataset. The adaptive time stepping scheme within mRM shows\na remarkable scalability compared to its predecessor. Median\nKling–Gupta efficiencies change less than 3\u2009% when the model\nparameterisation is transferred from 3 to 48\u2009km resolution. mRM also\nexhibits seamless scalability in time, providing similar results\nwhen forced with hourly and daily runoff. The streamflow calculated\nover the Danube catchment by the regional climate model REMO coupled\nto mRM reveals that the 50\u2009km simulation shows a smaller bias with\nrespect to observations than the simulation at 12\u2009km\nresolution. The mRM source code is freely available and highly\nmodular, facilitating easy internal coupling in existing Earth\nsystem models.