On the uncertainty in mesoscale modeling caused by surface parameters

Summary

The uncertainty of plant- and soil parameters (albedo, evaporative conductivity, roughness-length, soil volumetric heat capacity, field capacity, capillarity, emissivity), soil-type, subgrid-scale heterogeneity and inhomogeneity on mesoscale model results (e.g., fluxes, variables of state, cloud- and precipitation-formation) is explored. Simulations are performed wherein one parameter at a time is changed for all grid cells of the domain within its natural range of variability. The land-surface scheme employs the frequently operated force-restore-method and a one-dimensional heat-diffusion equation. Transpiration is considered by the often applied “big-leaf/big-stoma”-approach.

Except for wind, cloud- and precipitating-particles the uncertainty in the simulated quantities due to soil- or plant-parameters is lower at night and in the early afternoon than at the other time of the day. Therefore, evaluation of meteorological models should be performed at those times when the sensitivity to the choice of plant- and soil-parameters is at a minimum. According to the calculated probability density functions and similarity coefficients, the partitioning of excess water vapor between cloud water and ice can be significantly affected by the choice of field capacity, capillarity, volumetric heat capacity, thermal diffusivity and the emissivity of the earth’s surface. Considering these findings domain-specific and actual parameters should be preferred.