GC32C-03
Prospects for the Moon as an SI-Traceable Absolute Spectroradiometric Standard for Satellite Remote Sensing

Wednesday, 16 December 2015: 10:50
3014 (Moscone West)
Claire E Cramer1, Thomas C Stone2, Keith Lykke3 and John T Woodward1, (1)National Institute of Standards and Technology Gaithersburg, Gaithersburg, MD, United States, (2)US Geological Survey, Flagstaff, AZ, United States, (3)NIST, Gaithersburg, MD, United States
Abstract:
The Earth's Moon has many physical properties that make it suitable for use as a reference light source for radiometric calibration of remote sensing satellite instruments. Lunar calibration has been successfully applied to many imagers in orbit, including both MODIS instruments and NPP-VIIRS, using the USGS ROLO model to predict the reference exoatmospheric lunar irradiance. Sensor response trending was developed for SeaWIFS with a relative accuracy better than 0.1 % per year with lunar calibration techniques. However, the Moon rarely is used as an absolute reference for on-orbit calibration, primarily due to uncertainties in the ROLO model absolute scale of 5%-10%. But this limitation lies only with the models - the Moon itself is radiometrically stable, and development of a high-accuracy absolute lunar reference is inherently feasible. A program has been undertaken by NIST to collect absolute measurements of the lunar spectral irradiance with absolute accuracy <1 % (k=2), traceable to SI radiometric units. Initial Moon observations were acquired from the Whipple Observatory on Mt. Hopkins, Arizona, elevation 2367 meters, with continuous spectral coverage from 380 nm to 1040 nm at ~3 nm resolution. The lunar spectrometer acquired calibration measurements several times each observing night by pointing to a calibrated integrating sphere source. The lunar spectral irradiance at the top of the atmosphere was derived from a time series of ground-based measurements by a Langley analysis that incorporated measured atmospheric conditions and ROLO model predictions for the change in irradiance resulting from the changing Sun-Moon-Observer geometry throughout each night. Two nights were selected for further study. An extensive error analysis, which includes instrument calibration and atmospheric correction terms, shows a combined standard uncertainty under 1 % over most of the spectral range. Comparison of these two nights' spectral irradiance measurements with predictions from the ROLO model shows wavelength-dependent differences of up to several percent. We will report on these ongoing measurement and comparison efforts.