Shortwave clear-sky diffuse irradiance in the 300- to 1100-nm range: comparison of models with UV-VIS-NIR and broadband radiometer measurements at the Southern Great Plains ARM site in September/October 2001

Broadband shortwave diffuse horizontal irradiance models overestimate measurements by between 7 and 14% using the most reliable input data for the models and the best available broadband measurements of diffuse irradiance. This paper uses spectral irradiance measurements and models as opposed to broadband measurements and models to investigate the contributions to this difference from various regions of the spectrum. The data are from the first Atmospheric Radiation Measurement (ARM) diffuse irradiance intensive observation period (IOP) held in September and October of 2001 at the Oklahoma ARM site near Ponca City. Visible and ultraviolet (UV) rotating shadowband spectroradiometers (RSS) acquired data during the IOP. Diffuse measurements with conventional broadband diffuse pyranometers and direct irradiance measurements using an absolute cavity radiometer are also available for analysis. Integrated spectral measurements are consistent with broadband measurements and, therefore, confirm the earlier results that models over predict diffuse. The wavelength dependent differences in models and measurements are illustrated and discussed.