Evaluation of the forcing of climate by aerosol scattering of shortwave radiation in cloud-free conditions (direct aerosol forcing) requires knowledge of aerosol optical properties on relevant spatial and temporal scales. It is convenient to measure these properties at the surface. However, before these measurements can be used to quantitatively estimate direct climate forcing, it is necessary to determine the extent to which these properties are representative of the entire atmospheric column. In this paper we compare aerosol optical depth (AOD) determined by Sun photometry at the Southern Great Plains (SGP) Atmospheric Radiation Measurement (ARM) site in north central Oklahoma for several cloud-free days with estimates of AOD based on two methods. First, the aerosol extinction measured at the surface (taken as the sum of the aerosol scattering and absorption coefficients at instrumental relative humidity of∼20%) is multiplied by the mixing height determined from temperature profiles from radiosonde measurements. Even under conditions of vigorous midday mixing this approach underestimates AOD by as much as 70% using dry aerosol measurements and by roughly 40% when hygroscopic growth of aerosol under ambient relative humidity is taken into account. This discrepancy is attributed primarily to underestimation of aerosol column extinction, as confirmed by examination of normalized aerosol backscatter profiles obtained from micropulse lidar (MPL), which show substantial contributions of aerosol loading above the atmospheric boundary layer. The second approach uses MPL profiles of normalized aerosol backscatter to estimate the vertical profile of aerosol extinction using surface values. The resulting AOD's are on average 30% less than measured values. This discrepancy is attributed to hygroscopic growth of aerosols in the atmospheric column. The results show that at the SGP site even under conditions of vigorous mixing in the atmospheric boundary layer the aerosol optical depth cannot be estimated with surface measurements of aerosol extinction unless information on the vertical profile of aerosol extinction is taken into account.