The Impact of Land-Surface Parameter Properties and Resolution on the Simulated Cloud-Topped Atmospheric Boundary Layer

Abstract

Sensitivity tests using the ‘Consortium for Small Scale Modeling’ model in large-eddy simulation mode with a grid spacing of 100 m are performed to investigate the impact of the resolution of soil- and vegetation-related parameters on a cloud-topped boundary layer in a real-data environment. The reference simulation uses the highest land-surface parameter resolution available for operational purposes (300 m). The sensitivity experiments were conducted using spatial averaging of about $$2.5\,\hbox {km}\times 2.5\,\hbox {km}$$ and $$10\,\hbox {km} \times 10\,\hbox {km}$$ for the land-surface parameters and a completely homogeneous distribution for the whole model domain of about $$70\,\hbox {km} \times 70\,\hbox {km}$$ . Additionally, one experiment with a higher mean soil moisture and another with six mesoscale patches of enhanced or reduced soil moisture are performed. Boundary-layer clouds developed in all simulations. To assess the deviations of cloud cover on different scales within the model domain, we calculated the root-mean-square deviation (RMSD) between the sensitivity experiments and the reference simulation. The RMSD depends strongly on the spatial resolution at which cloud fields are compared. Different spatial resolutions of the cloud fields were generated by applying a low-pass filter. For all sensitivity experiments, large RMSD values occur for cut-off wavelengths $${<}1$$ km, reflecting the stochastic nature of convection, but they decrease rapidly for wavelengths between 1 and 5 km. For cut-off wavelengths $${>}5\,\hbox {km}$$ , the RMSD is still pronounced for the simulation with higher mean soil moisture. Additionally, for cut-off wavelengths between 5 and 30 km, considerable differences can be found for the experiment with mesoscale patches and for that with homogeneous land-surface parameters. Spatial averaging of land-surface parameters for areas of $$2.5\,\hbox {km} \times 2.5\,\hbox {km}$$ and $$10\,\hbox {km} \times 10\,\hbox {km}$$ results in larger patch sizes but simultaneously in reduced amplitudes of land-surface parameter anomalies and shows the lowest RMSD for all cut-off wavelengths.