The daily surface and atmospheric radiative fluxes from NASA Clouds and the Earth's Radiant Energy System (CERES) Synoptic 1 degree (SYN1deg) Ed3A are among the most widely used data to study cloud‐radiative feedback. The CERES SYN1deg data are based on Fu‐Liou radiative transfer computations that use specific humidity ( Q ) and air temperature ( T ) from NASA Global Modeling and Assimilation Office (GMAO) reanalyses as inputs and are therefore subject to the quality of those fields. This study uses in situ Q and T observations collected during the Dynamics of the Madden‐Julian Oscillation (DYNAMO) field campaign to augment the input stream used in the NASA GMAO reanalysis and assess the impact on the CERES daily surface and atmospheric longwave estimates. The results show that the assimilation of DYNAMO observations considerably improves the vertical profiles of analyzed Q and T over and near DYNAMO stations by moistening and warming the lower troposphere and upper troposphere and drying and cooling the mid‐upper troposphere. As a result of these changes in Q and T , the computed CERES daily surface downward longwave flux increases by about 5 W m⁻², due mainly to the warming and moistening in the lower troposphere; the computed daily top‐of‐atmosphere (TOA) outgoing longwave radiation increases by 2–3 W m⁻² during dry periods only. Correspondingly, the estimated local atmospheric longwave radiative cooling enhances by about 5 W m⁻² (7–8 W m⁻²) during wet (dry) periods. These changes reduce the bias in the CERES SYN1deg‐like daily longwave estimates at both the TOA and surface and represent an improvement over the DYNAMO region.