Comparative Analysis of Radiance-based and Reflectance-based Reflectance Retrievals

Early earth science imagers in the reflected solar part of the spectrum relied on a radiance-based SI-traceability path to top-of-atmosphere reflectance. More recent missions tend to use reflectance-based approaches as space-grade diffusers became better understood. The difference between the two approaches will become more important for next generation missions where low-cost, low-mass sensors shift back to a radiance-based approach and lower absolute uncertainties are becoming necessary. Radiance-based retrievals are so named because they rely on the absolute radio-metric calibration of the sensor to determine scene spectral radiance that is converted to reflectance using an assumed solar spectral irradiance model. Reflectance-based approaches convert measurements directly to reflectance through ratios to views of a diffuser reference of known reflectance. The current work describes the use of a field deployable transfer radiometer as well as commercial field spectrometer data to draw out the discrepancies rooted in differences from the SI-traceability in order to understand the limits for harmonizing multi mission data. The results show that surface reflectance results calculated using the two retrievals differ even when identical radiometer signals are used. The ground-based retrievals are complicated by the presence of a diffuse sky irradiance, but the use of the vicarious calibration site at Railroad Valley Playa and the surface areas selected for this study help to limit those effects. Sensitivity studies for the retrievals are also presented to highlight the traceability-based differences rather than those caused by the collection and processing of the data. The comparison between the two traceability paths shows clear differences that are within the combined uncertainties of the two approaches but with consistent systematic effects.