Forecasts from an operational system for predicting global diurnal sea surface temperature (SST) variability are compared against multichannel sea surface temperatures calculated from advanced very high resolution radiometer data from the NOAA-9 satellite. The forecast system consists of an upper-ocean model driven by heat fluxes and wind stresses from an operational global weather prediction model. Visual pattern correlations between observed and forecast diurnal changes in SST are generally favourable, indicating that the forecast system is useful. Some systematic deficiencies are also noted: the forecasts do not reproduce the surprisingly sharp boundaries of diurnal warming regions and underpredict the amplitude of the extreme warming events