Item

Abstract

Forecasting high impact weather events is a major challenge for numerical weather prediction. Initial condition uncertainty plays an important role but so do potentially uncertainties arising from the representation of subgrid-scale processes, e.g. cloud microphysics. In this project, we investigate the impact of cloud microphysical parameter uncertainties on the forecast of a selected severe convective storm over South-Eastern Germany in 2019. A perturbed parameter ensemble with 112 members has been generated with the ICON model (2-moment cloud microphysics, 1 km grid-spacing). Uncertain parameters in the representation of (i) CCN and INP activation, (ii) diffusional growth of ice and snow, (iii) riming of graupel and hail, as well as (iv) the mass-diameter and mass-fall velocity relations for graupel and hail are considered. To generate systematic parameter variations from the eight-dimensional parameter space a latin hypercube sampling is used. The storm properties, including cloud microphysical structure and surface (hail) precipitation, are determined by identifying storm objects with a storm tracking and watershedding algorithm. Results indicate a large impact of INP concentration on precipitation and surface hail amounts. Furthermore, a strong interaction of INP and CCN concentrations is found with increased sensitivity to INP concentrations at low CCN concentrations. Statistical emulation and variance-based sensitivity analysis further indicate substantial impact of graupel density and ice capacitance on surface precipitation and hail content respectively. Additionally, an initial condition ensemble is simulated. First insights into the relative importance of initial condition uncertainty and microphysical parameter uncertainty will be presented.