Estimates from semiempirical models that characterize surface heat flux, mixing depth, and profiles of temperature, wind, and turbulence are compared with observations from atmospheric field studies conducted in Colorado, Illinois, Indiana, and Minnestoa. In addition, for wind and turbulence profiles, sodar observations are compared with tower measurements at the Colorado site. The median surface heat flux, as calculated using surface-layer flux-profile relationships and an energy budget model, was consistently overestimated from 20% to 80%. Wind profiles were derived using surface-layer flux-profile relationships, a wind-profile power-law based on Pasquill stability category, and sodar measurements. The sodar measurements were superior to both types of model estimates. Turbulence profiles were derived from sodar measurements and from semiempirical similarity relationships based on mixing depth and Obukhov length. The scatter in the comparisons with the sodar observations is twice that seen in the comparisons with empirical profile relationships. Overall, it appears that uncertainty of as low as 20% to 30% in the characterization of the diffusion meteorology is the exception rather than the rule.