On the sensitivity of bow-echo ensemble forecasts to grid spacing

The bow echo, a mesoscale convective system responsible for much hail and wind damage across the United States, is associated with poor skill in convective-mode forecasts. Previous studies have shown great sensitivity of their forecasts to model error, such as perturbations from the stochastic kinetic-energy backscatter (SKEB) scheme, and  changes in microphysical parameterization. In addition, the increase in grid resolution across numerous operational weather centers may change the nature of error introduced by the smaller scale of truncation. Here, we investigate the effect of increasing resolution within limited-area ensemble forecasts in the central United States during two strong bow-echo events.  The signal-to-noise ratio for each case and grid resolution is increased through use of SKEB perturbations that create equally likely ensemble members.

We may expect finer resolutions to have higher skill due to a more faithful, explicit reproduction of atmospheric processes, and this skill is evaluated with an adaptation of the object-based SAL verification method. We also expect larger spread of convective modes due to the capture of more small-scale variance. Our results show that both ensemble spread and skill of the finer resolution ensemble is similar or lower than the coarser resolution simulations. Moreover, bow-echo system speed is notably faster when grid resolution increases. We propose this may be due to the increased (fractal) cloud surface area in higher resolutions, which allows greater entrainment of dry air and hence increased evaporative cooling.