Abstract. Radiative forcing provides an important basis for understanding and predicting global climate changes, but its quantification has historically been done independently for different forcing agents, involved observations to varying degrees, and studies have not always included a detailed analysis of uncertainties. The Copernicus Atmosphere Monitoring Service reanalysis is an optimal combination of modelling and observations of atmospheric composition. It provides a unique opportunity to rely on observations to quantify the monthly- and spatially-resolved global distributions of radiative forcing consistently for six of the largest forcing agents: carbon dioxide, methane, tropospheric ozone, stratospheric ozone, aerosol-radiation interactions, and aerosol-cloud interactions. These radiative forcing estimates account for adjustments in stratospheric temperatures, but do not account for rapid adjustments in the troposphere. On a global average and over the period 2003–2016, stratospherically adjusted radiative forcing of carbon dioxide has averaged +1.84 W m$^{−2}$ (5–95% confidence interval: 1.46 to 2.22 W m$^{−2}$) relative to 1750 and increased at a rate of 17 % per decade. The corresponding values for methane are +0.45 (0.35 to 0.55) W m$^{−2}$ and 3 % per decade, but with a clear acceleration since 2007. Ozone radiative forcing averages +0.32 (0 to 0.64) W m$^{−2}$ and aerosol radiative forcing averages −1.37 (−2.17 to −0.57) W m$^{−2}$. Both have been relatively stable since 2003. Taking the six forcing agents together, there no indication of a slowdown or acceleration in the rate of increase in anthropogenic radiative forcing over the period. These ongoing radiative forcing estimates will monitor the impact on the Earth’s energy budget of the dramatic emission reductions towards net-zero that are needed to limit surface temperature warming to the Paris Agreement temperature targets. Indeed, such impacts should be clearly manifested in radiative forcing before being clear in the temperature record. In addition, this radiative forcing dataset can provide the input distributions needed by researchers involved in monitoring of climate change, detection and attribution, interannual to decadal prediction, and integrated assessment modelling. The data generated by this work are available at https://doi.org/10.24380/ads.1hj3y896 (Bellouin et al., 2020).