A Stochastic Perturbed Parameterization Tendency Scheme for Diffusion (SPPTD) and Its Application to an Idealized Supercell Simulation

Abstract

iffusion plays an important role in supercell simulations. A stochastically perturbed parameterization tendency scheme for diffusion (SPPTD) is developed to incorporate diffusive uncertainties in ensemble forecasts. This scheme follows the same procedure as the previously published stochastically perturbed parameterization tendencies (SPPT) scheme but uses a recursive filter to generate smooth perturbations. It also employs horizontal and vertical localization to retain the impact of perturbation in areas with strong shear. Three additional restrictions are added for the sake of integration stability; these restrictions determine the area and amplitude of the perturbation and the situation to suspend SPPTD.The performance of this scheme is examined by using an idealized supercell storm. The model errors are simulated using different resolutions in the truth run (1 km) and ensemble forecasts (2 km). The results indicate that the ensemble forecasts using SPPTD encompass the intensity and displacement of maximum updraft helicity in the truth run. This scheme yields better results than can be obtained using initial perturbations or larger computational mixing coefficients.The sensitivity of SPPTD to each of its parameters is also examined. The results indicate that the optimal horizontal and temporal scales for SPPTD are 40 km and 30 min, respectively. Moderately adjusting the spatiotemporal scale by 10 km or 10 min does not significantly change the SPPTD performance. In this case study, an ensemble size of 20 is sufficient. Perturbing the diffusion terms of all variables using the same approach does not provide additional benefits other than that of selected variables and thus requires further study.