Sensitivity to the cloud microphysics scheme of the simulation of orographic precipitation

There has been a considerable increase in computer power in recent years allowing for the use of more sophisticated cloud microphysics schemes in high-resolution atmospheric models. Despite this, fully explicit bin-resolving schemes are generally prohibitively expensive for 3D simulations. Consequently, bulk microphysics schemes (BMSs) continue to play an important role in both research and in operational NWP. Several BMSs have been developed in the last decade which predict two or more moments of the hydrometeor size distributions. There has, however, been relatively little investigation of the benefits of using higher-moment over single-moment schemes.

In this work, we continue to examine the utility of the multi-moment BMS developed by Milbrandt and Yau (2005a,b). During November-December 2001, the IMPROVE-2 campaign collected a comprehensive set of microphysical data over the Oregon Cascades with the objective of diagnosing and correcting problems associated with current BMSs. This experiment provides a useful data set for testing and calibrating the multi-moment scheme in the context of a case of orographically enhanced precipitation. Simulations of the 13-14 December 2001 case with a mesoscale model at grid-spacings of 4 km and 1 km have been performed using various versions of the multi-moment scheme.

The full triple-moment version was fairly successful at simulating the observed precipitation over a region of complex terrain. While there were little differences between the triple-moment and double-moment simulations, there were notable, though not dramatic, differences in the single-moment simulation. This is contrast with a similar study using the multi-moment scheme to simulate deep convection, where much greater sensitivity to the number of predicted moments was found. However, there were large differences between the simulation of the current case using the single-moment version of the scheme and a simulation using a different single-moment scheme. For this case of orographic precipitation, it appears that the way the ice-phase hydrometeor spectrum is partitioned into categories in the scheme is at least as important as the number of moments predicted.