Impact of combined uncertainties on convective precipitation during different synoptic control

Subgrid-scale uncertainty for convective-scale forecasts and its sensitivity to synoptic conditions remain unclear. In this study, the relative importance of individual and combined perturbations representing various sources of uncertainties are addressed by simulating real case studies stratified into weak and strong synoptic convective-forcing situations using the quasi-operational setup of ICON-D2 model with the double moment microphysics scheme. We quantified and intercompared the impact given by three uncertainties, microphysics parameters (cloud condensation nuclei concentrations and shape parameters of cloud droplet size distribution), initial and boundary conditions (IBC), and physically-based stochastic perturbation scheme (PSP). The impact on two aspects of precipitation forecasts are assessed: daily accumulated precipitation amount averaged over Germany and temporal evolution of spatial variability of hourly precipitation. 
The relative impact on 24h precipitation is quantified against the mean of sub-ensembles, in which different parameters are perturbed. Microphysics perturbations and PSP show a minor impact compared to the impact of IBC perturbations, but combining IBC and microphysics perturbations extends the tails of the forecast distribution.  The responses to all the perturbations are relatively insensitive for strong-forcing situations.
Visual inspection of hourly precipitation fields reveals that microphysics perturbations slightly shift convective cells but vary precipitation intensities while IBC perturbations completely deform the spatial convective structure. We attempted to quantify the typical scales of modification using Fractions Skill Scores and found that microphysics perturbations have an explicit impact on the position of individual convective cells. On the other hand, PSP initiates convection earlier and affect the lifetime of cloud and convective precipitation, resulting in earlier and faster growth of spatial variability of hourly precipitation than microphysics. 
In the talk, additional comparisons against radar observations and benefits of practical forecast skills will be also presented.