Sensitivity of Convection-Allowing Forecasts to Land-Surface Model Perturbations and Implications for Ensemble Design

In this exploratory study, a series of perturbations to the land-surface model (LSM) component of the Weather Research and Forecasting model was developed to investigate the sensitivity of forecasts of severe thunderstorms and heavy precipitation at 4 km grid spacing. The perturbations were applied to a 10-member ensemble including other mixed physics parameterizations and compared against an identically configured ensemble that did not include the LSM perturbations to determine their impact on probabilistic forecasts. A third ensemble using only the LSM perturbations was also configured. A case involving tornadic supercells evolving upscale to a mesoscale convective system is detailed to illustrate the physical impact of the LSM perturbations. Statistical results from 14 other cases featuring intense convection with heavy precipitation are also presented.

The case study revealed a wide variety of behaviors can be induced on the mesoscale by the LSM perturbations, especially in terms of boundary layer heat and moisture content. Large variations in sensible and latent heat flux caused by the perturbations resulted in ensemble diversity rivaling that of an ensemble using other commonly used methods of perturbing model physics. The case study results are consistent with the statistical analysis showing increased ensemble spread and decreased error when incorporating the LSM perturbations. Probabilistic precipitation forecasts were also improved.

The results of this study suggest LSM perturbations can sample a dimension of model error not yet sampled by other methods and should be included in convection-allowing ensembles.