Observations and simulations of mesoscale convective systems

Mesoscale convective systems (MCSs) are studied herein using observations and numerical weather prediction models. Observations from the bow-echo and mesoscale convective vortex experiment (BAMEX) were analyzed and compared with numerical simulations of MCSs that use parameterized and explicit representations of convection. Unfortunately a direct comparison was not possible because the models used here could not reproduce observed MCS structures. Thus we examined how model physical parameterizations influence representations of convection in simulations using explicit and parameterized convection.; The analysis of observed vertical profiles of temperature and moisture from BAMEX revealed spatio-temporal variability of temperature, moisture, and wind within MCSs. We hypothesize that this was likely due to the spatio-temporal variability of microphysical processes such as melting, sublimation and evaporation of hydrometeors. We found that vertical profiles of wet bulb potential temperature implied that air from a variety of source regions merged to form the environment underneath MCSs characterized by mesoscale unsaturated downdrafts. The layered profiles may be a result of the spatio-temporal variability of microphysical, or dynamical, or a combination of both processes.; A model analysis of 2D idealized MCS structure and propagation using a convective parameterization scheme (CPS) was performed. The main findings indicated that the lack of an evolving temperature and moisture adjustment impeded the development of grid-resolved mesoscale circulations including the major flow branches of MCSs and the development of the cold pool. Furthermore, sounding structures produced by the model lacked the microphysical effects such as melting and evaporation which characterize the stratiform rain region. We found that gravity waves were mostly responsible for initiating and sustaining parameterized convection. Sensitivity tests revealed that the depth of cold pool and heating rates controlled the propagation speeds of parameterized convection. Furthermore, scheme modifications proved insufficient to significantly alter propagation speeds relative to the control simulation.; The Weather Research and Forecast model was used to examine the planetary boundary layer (PBL) scheme dependence on convective initiation using an explicit (near cloud resolving) and implicit (parameterized) convection approach. Initiation and propagation of the simulated convection were compared and contrasted. Systematic warm biases of virtual potential temperature and stronger vertical velocity were present when comparing the Mellor-Yamada-Janjic (MYJ) and Yonsei University (YSU) PBL schemes. These biases led to different convective initiation and organization of the MCS. The vertical velocity bias was reduced when using a CPS but the temperature and moisture bias was still present and of the same magnitude as the explicit simulations. The simulations using a CPS initiated convection early relative to the explicit simulations while subsequent initiation was perpendicular to frontal boundaries, mimicking propagation. The explicit simulations of MCSs were dependent upon PBL scheme design with the MYJ scheme producing realistic mesoscale vortices along the broken gust front. The YSU gust front was very uniform in terms of vorticity and convergence and, consequently, longer lived leading to long lived MCSs.; We conclude that comparison between numerical simulations and observations requires resolution on the order of hundreds of meters to account for the observed spatio-temporal variability of temperature and moisture. Only recently has this been achieved in idealized modeling of MCSs and thus appears as a distant goal. However the fine scale simulations performed here will suffice for operational numerical weather prediction if the results shown here apply to other cases. In the mean time, it appears reasonable to improve convective parameterization for regional and glo...