The design, construction, and calibration of a spectral diffuse/global irradiance meter

Vicarious calibration methods have been developed to calibrate radiometric sensors in-flight. One such method, the irradiance-based method, requires the measurement of the diffuse-to-global (diffuse-to-total) irradiance ratio. Diffuse/global irradiance measurements may also be used to deduce atmospheric descriptors and provide a comparison with atmospheric modeling predictions. I describe the design, construction, calibration, and application of a spectral diffuse/global irradiance meter that can accomplish these objectives in this dissertation. I develop general integrating sphere theory, modeling methods, and describe the resultant computer model. The model results agreed with theory to better than 1% for a simple unbaffled integrating sphere. I applied the model to design an interior baffled integrating sphere-based cosine collector. I developed a method of tolerating the thermal expansion of Spectralon® and the collector was constructed. Measurements of the collector angular response agreed with the model predictions to better than 4% for input zenith angles from 0° to 70°. The resulting instrument is automated and collects diffuse and global irradiance from 300 nm to 1100 nm. It has a nominal 12 nm full-width at half-maximum bandpass and has a minimum sampling interval of 1 nm. I estimate the uncertainty of the measurements to be 3.2%. The largest contributor to the total uncertainty is the measurement uncertainty of the diffuse irradiance at 2.5%. The instrument was used in a field experiment. Optical depths derived from the diffuse/global irradiance measurements agreed with those derived from a solar radiometer to within 0.008. Diffuse-to-global irradiance measurements made by the instrument were compared with an independent method and found to generally agree within 6%. The measurements were consistently lower than radiative transfer modeling estimates. Top of the atmosphere relative radiances computed from the two independent diffuse-to-global irradiance data sets generally agreed to better than the 2.9% uncertainty associated with the diffuse/global irradiance meter data set. The relative radiances of the diffuse/global data set collected with the instrument were within the 4.9% uncertainty estimate of the reflectance-based method