Presentation and discussion of the high-resolution atmosphere–land-surface–subsurface simulation dataset of the simulated Neckar catchment for the period 2007–2015

Abstract

Abstract. Coupled numerical models, which simulate water and energy fluxes in the subsurface–land-surface–atmosphere system in a physically consistent way, are a prerequisite for the analysis and a better understanding of heat and matter exchange fluxes at compartmental boundaries and interdependencies of states across these boundaries. Complete state evolutions generated by such models may be regarded as a proxy of the real world, provided they are run at sufficiently high resolution and incorporate the most important processes. Such a simulated reality can be used to test hypotheses on the functioning of the coupled terrestrial system. Coupled simulation systems, however, face severe problems caused by the vastly different scales of the processes acting in and between the compartments of the terrestrial system, which also hinders comprehensive tests of their realism. We used the Terrestrial Systems Modeling Platform (TerrSysMP), which couples the meteorological Consortium for Small-scale Modeling (COSMO) model, the land-surface Community Land Model (CLM), and the subsurface ParFlow model, to generate a simulated catchment for a regional terrestrial system mimicking the Neckar catchment in southwest Germany, the virtual Neckar catchment. Simulations for this catchment are made for the period 2007–2015 and at a spatial resolution of 400 m for the land surface and subsurface and 1.1 km for the atmosphere. Among a discussion of modeling challenges, the model performance is evaluated based on observations covering several variables of the water cycle. We find that the simulated catchment behaves in many aspects quite close to observations of the real Neckar catchment, e.g., concerning atmospheric boundary-layer height, precipitation, and runoff. But also discrepancies become apparent, both in the ability of the model to correctly simulate some processes which still need improvement, such as overland flow, and in the realism of some observation operators like the satellite-based soil moisture sensors. The whole raw dataset is available for interested users. The dataset described here is available via the CERA database (Schalge et al., 2020): https://doi.org/10.26050/WDCC/Neckar_VCS_v1.

Assimilation of Groundwater Level and Soil Moisture Data in an Integrated Land Surface‐Subsurface Model for Southwestern Germany

“…Furthermore, a 3D geologic model of the geological survey of Baden-Württemberg defined by 11 rock types was used, which should provide a realistic representation of the deeper subsurface of the domain. Using this setting, Schalge et al (2021) demonstrated that the simulated GWL and river discharge were in good correspondence with the observed values.…”

“…Using this setting, Schalge et al. (2021) demonstrated that the simulated GWL and river discharge were in good correspondence with the observed values.…”

“…The DA experiments in this work are carried out with TSMP‐PDAF, which includes a coupled land surface‐subsurface model and a DA framework (Kurtz et al., 2016; Nerger & Hiller, 2013; Shrestha et al., 2014). Measurements are extracted from a virtual reality (VR) simulation (hereafter we name it as VR) which mimics southwestern Germany, including the Neckar River catchment (Schalge et al., 2021). The simulation with the coupled model is based on realistic atmospheric forcings, land use including leaf area indices, soil maps and geological maps from this region in southwestern Germany.…”

An elastic framework for ensemble-based large-scale data assimilation

“…For our test, we were provided observations from 25 ground water measuring sensors distributed over the whole catchment. Observation values were taken from a virtual reality simulation (Schalge et al, 2020). ParFlow runners are parallelized on one full node (40 processes for experiments on Jean-Zay and 48 processes for JUWELS, respectively).…”

Towards hourly three-dimensional ensemble data assimilation of screen-level observations into coupled atmosphere-land models

“…The simulations span a region of 267 km × 302 km in longitudinal and latitudinal direction, respectively, located in the south-western part of Germany with the embedded Neckar catchment. As similarly used in Schalge et al (2021), the heterogeneous land surface setup typifies conditions in mid-latitudes and Central Europe. The lateral boundary conditions in the atmosphere are based upon one member of the COSMO-DE-EPS ensemble (Peralta et al, 2012) with a similar configuration to our atmospheric setup.…”