Direct estimates of surface radiative fluxes that resolve regional and weather-scale variability over the whole globe with reasonable accuracy have only become possible with the advent of extensive global, mostly satellite, data sets within the past couple of decades. The accuracy of these fluxes, estimated to be about 10--15 W/m2, is largely limited by the uncertainties of the input data sets. This study presents a fuller, more quantitative evaluation of these uncertainties, mainly for the near-surface air temperature and humidity, by comparing the main available global data sets from the European Centre for Medium-Range Weather Forecasts, NASA, the National Centers for Environmental Prediction, the International Satellite Cloud Climatology Project (ISCCP) and the Laboratoire de M¿t¿orologie Dynamique that are treated as ensemble realizations of actual climate such that their differences represent an estimate of the uncertainty in their measurements (because we do not know the absolute truth). The results are globally representative and may be taken as a generalization of our previous ISCCP-based uncertainty estimates for the input data sets. Near-surface atmospheric properties have the primary role in determining the surface downward longwave (LW) flux. From this study, the most important quantity, the surface air temperature, has a uncertainty of about 2--4 K (3 K on average), which would easily induce about 15 W/m2 uncertainty for surface downward LW flux. The humidity profile comparison suggests an uncertainty of 20--25% for the atmospheric column precipitable water below the 300 hPa level, which would cause $lesssim$10 W/m2 uncertainty for surface downward LW flux, making it the second largest source of uncertainty. The comparison for the difference between surface skin and air temperature shows its uncertainty is about 2--3 K, which translates into 10--15 W/m2 uncertainty for surface net LW flux. The used atmospheric data set from ISCCP represents the diurnal variations better than the other available sources (as it was designed to do) and the synoptic variations only slightly better than the other sources, but it still has notable clear-cloudy sky biases and interannual variations that are dominated by spurious changes introduced by methodology changes in the original TOVS product. In a companion paper, the work is extended to evaluate the uncertainties of surface radiative properties.