Three-dimensional week-long simulations of TOGA COARE convective systems using the MM5 mesoscale model

A three-dimensional nonhydrostatic mesoscale model, the Pennsylvania State University/National Center for Atmospheric Research mesoscale model (MM5), is used to simulate the evolution of convective systems over the intensive flux array (IFA) during the Tropical Ocean Global Atmosphere Coupled Ocean- Atmosphere Response Experiment, during 19-26 December 1992. The model is dr iven by a time-varying "IFA mean forcing" based on the average advective te ndencies of temperature and moisture over the IFA. The domain-averaged hori zontal wind is kept close to the observed IFA mean using Newtonian relaxati on. Periodic lateral boundary conditions are imposed. Simulations with thre e horizontal grid spacings, 2, 15, and 60 km, are conducted. With 15- and 6 0-km resolution, subgrid-scale cumulus convection is parameterized while me soscale convective organization is explicitly resolved over a (600 km)(2) d omain. With 2-km resolution. convection is fully resolved over a (210 km)(2 ) domain. Despite their different horizontal resolution and different treatment of mo ist convection, the simulations all produce very similar temporal variabili ty in domain-averaged temperature and relative humidity profiles. They also closely resemble each other in various statistical properties of convectiv e systems. A comprehensive comparison of the 15- and 2-km model results aga inst observations is performed. The domain-averaged cloud amount and precip itation agree well with observations. Some shortcomings are noted. During s uppressed convective periods, the model tends to have greater areal coverag e of rainfall and more cirrus anvil clouds than observed. Over the 8-day pe riod, both models produce mean temperature drifts about 2 K colder than obs erved. A histogram of modeled cloud-top temperature captures the observed b reaks between convective episodes but shows excessive and persistent cold c irrus clouds. A radar reflectivity histogram shows that the 15-km model sli ghtly overpredicts radar reflectivity and that the 2-km model has too high and temporally homogeneous reflectivities. The model-simulated cloud cluste r size is somewhat smaller than the observed. Surface sensible and latent h eat fluxes are overestimated by 50%-100%, due both to shortcomings in the s urface Bur calculations in the model and model-produced mean temperature an d humidity biases. Downwelling solar flux at the surface is underestimated mainly because of the simple shortwave radiation scheme. This study suggests that large-domain simulations using the MM5 with 15-km resolution can be a useful tool for further study of tropical convective or ganization and its interaction with large-scale circulation.