Solar reflectance measurements (0.6 and 3.9 μm) from 15-min Geostationary Operational Environmental Satellite (GOES) 9 imager data were used to estimate cloud liquid water path (LWP) for an extensive stratocumulus system over Oklahoma on May 2, 1996. The objective was to determine the consistency between these satellite estimates and retrievals from high temporal resolution (20 s) surface microwave radiometer (SMWR) measurements. The SMWRs were located at the Atmospheric Radiation Measurement (ARM) program cloud and radiation test bed (CART) sites at Morris and Purcell in Oklahoma. Results show that while the comparisons are in favorable agreement at both sites in the morning and early afternoon (root-mean-square difference of 17 g m-2 and correlation of 0.94), large cloud LWP maxima in the midafternoon as measured by the SMWR at the Morris site are not captured by the satellite retrievals. On the basis of indirect evidence (in situ microphysical measurements were unavailable), it is hypothesized that the discrepancies may be the result of the formation of light drizzle in the middle to lower portions of the cloud, unseen at visible and near-infrared wavelengths from space but easily sensed by microwave radiometry. These results demonstrate that extra care must be taken in future efforts to validate satellite derived cloud properties on a routine basis using SMWR data. Additional information about the cloud microphysical properties may also be required to help properly interpret the comparisons, particularly in the later stages of development of stratocumulus. ¿ 1999 American Geophysical Union