Radiative forcing associated with land use change is largely derived from global circulation models (GCM), and the accuracy of these estimates depends on the robustness of the vegetation characterization used in the GCMs. In this study, we use observations from the Clouds and Earth's Radiant Energy System (CERES) instrument on board the Terra satellite to report top-of-the-atmosphere (TOA) radiative forcing values associated with clearing of native vegetation for agricultural purposes in southwest Australia. Over agricultural areas, observations show consistently higher shortwave fluxes at the TOA compared to native vegetation, especially during the time period between harvest and planting. Estimates using CERES observations show that over a specific area originally covered by native vegetation, replacement of half the area by croplands results in a diurnally averaged shortwave radiative forcing of approximately −7 W m⁻². GCM-derived estimates for areas with 30% or more croplands range from −1 to −2 W m⁻² compared to observational estimate of −4.2 W m⁻², thus significantly underestimating radiative forcing due to land use change by a factor of 2 or more. Two potential reasons for this underestimation are incorrect specification of the multiyear land use change scenario and the inaccurate prescription of seasonal cycles of crops in GCMs.