The diurnal cycle of the atmospheric boundary layer (ABL) in California's Central Valley is controlled by meteorological processes on various scales. Errors in model simulated ABL, as revealed by comparisons of numerical model simulations with observations indicate that there are errors in those meteorological processes. In this study, the 5-day averaged ABL diurnal evolution simulated by the NCAR/Penn State community mesoscale model (MM5) is systematically validated and evaluated using the observations from the Central California Ozone Study (CCOS) 2000 field experiment. The model simulated ABL winds and thermal structure, are compared with the observational counterparts. It is found that the ABL mixing parameterization in MM5 can simulate the ABL structure and dynamics in California's Central Valley reasonably well under clear sky when the large-scale, upper-level forcing and surface characteristics are specified correctly. Our evaluation indicates that the errors in the model simulations of ABL structure and dynamics are closely associated with errors in the model's upper level synoptic scale flow and uncertainties in the surface characteristics.