We present an improved self-adaptive methodology for the continuous estimation of downwelling clear-sky longwave (LW) radiative flux based on analysis of surface irradiance, air temperature, and humidity measurements that includes a term to account for near surface optically thin haze. Comparison between our estimations and clear-sky LW measurements for many years of data from the Atmospheric Radiation Measurement (ARM) Climate Research Facility’s Southern Great Plains (SGP), Tropical Western Pacific (TWP), and North Slope of Alaska (NSA) sites show agreement at about the 4 W m−2 level, with 75%, 94%, and 68% of the data falling within that range for the SGP, TWP, and NSA sites, respectively. Although there is no exact means of determining the uncertainty associated with the clear-sky LW estimations, our analyses and comparison with detailed radiative transfer (RT) model calculations suggest our estimations on average are no worse than model calculations that require temporally and spatially averaged input information. Our technique exhibits a high degree of repeatability for the downwelling LW cloud effect, with agreement at about the 3 W m−2 level. Applying our technique and that of Long and Ackerman (2000) to 15 years of data from the ARM SGP site shows the maximum all-sky and clear-sky SW and LW occurs during summer, with the greatest year-to-year clear-sky SW variability occurring in fall. The downwelling LW cloud effect is fairly constant across the seasons, but the greatest SW cloud effect occurs in spring. The downwelling net cloud effect is dominated by the SW, with the largest effect occurring in spring (−64 W m−2) and the smallest occurring during winter (−21 W m−2).