The project to intercompare regional climate simulations (PIRCS) is a community based effort that provides a common test bed for the regional climate modeling community. More than 15 regional climate models (RCMs) have completed a 60-day integration, the PIRCS 1b experiment. The output collected at Iowa State University allows us to perform a number of process intercomparisons among the models and with observations. The PIRCS 1b experiment covers 1993 central U.S. intense flood period which is composed of a series of heavy rainfall events associated with mesoscale convective systems. Precipitation parameterization in the models is a key to the success of the individual model simulations. This paper analyzes mesoscale signals in forcing mesoscale precipitation as parameterized in those models.

As constrained by common lateral boundary conditions, most RCMs reproduced the main large-scale features of the period. On the other hand, the models showed significant inter-model differences in mesoscale disturbance intensity and distribution. In spite of large diversities in mesoscale signals in the models, however, the resulting precipitation as computed via parameterization modules, mainly cumulus parameterization schemes, showed general similarity among models. This disconnection between mesoscale features and resultant precipitation reflects somewhat the decoupling between model (three dimensional) dynamics and (one-dimensional) thermodynamics in the models. This result also implies that a RCM might produce correct precipitation while mishandling mesoscale dynamics, signaling the need for detailed analysis of model internal dynamics and not just rainfall on the ground.