The atmospheric correction of hyperspectral imaging data remains the largest source of uncertainty in the derivation of surface reflectance and the data products created from surface reflectance.…

This contract was awarded by the National Aeronautics and Space Administration (NASA) Shared Services Center to The Regents of the University of Colorado, doing business as the University of Colorado Office of Contracts and Grants Division. The $154,515.50 grant funds research to improve atmospheric correction of hyperspectral imaging data, focusing on the aerosol component of radiative transfer modeling. The goal is to generate a library of aerosol spectra based on AVIRIS observations and MODIS data to enhance atmospheric correction for current and future hyperspectral imaging over India, a region with complex aerosol conditions. This is intended to reduce a key source of uncertainty in deriving surface reflectance from AVIRIS data. The University of Colorado, an accredited public research institution, will apply its expertise in atmospheric science, remote sensing, and optimal estimation methodologies to achieve these research objectives. No set-aside designation was used for this contract award.

The atmospheric correction of hyperspectral imaging data remains the largest source of uncertainty in the derivation of surface reflectance and the data products created from surface reflectance. Despite the dramatic gains in signal-to-noise ratio of the aviris instrument over the past two decades, the atmospheric radiative transfer modeling required to convert radiance to reflectance remains the greatest source of error because the atmospheric correction has not kept pace with the instrumental improvements. From an atmospheric correction standpoint, the aviris dataset collected over india presents a unique opportunity. India is a very challenging environment in terms of the aerosol environment encountered by previous aviris and future hyperspectral observations. India routinely reports large aerosol optical thicknesses and also contains an especially rich environment of different aerosol types. This makes the improvement of atmospheric correction critical to the future of hyperspectral imaging data in india. This proposal seeks to improve on atmospheric correction of imaging spectrometer data, focusing on the aerosol component of atmospheric radiative transfer modeling. Currently, nearly all of the standard atmospheric correction routines (e.g. atcor, flaash, etc) use the modtran radiative transfer model as the basis for the calculation of the terms required for surface reflectance retrieval. The default aerosol types are restricted to very few aerosol types (e.g. urban , desert , rural , etc.) this proposal will generate a library of aerosol spectra (i.e. an aviris-india aerosol spectral catalog) based on a combination of aerosol properties retrieved from aviris observations in 2015 and 2016, and aerosol properties measured by the moderate resolution imaging spectroradiometer (modis) experiment, aided by observations made within india from ground-based retrievals of aerosol optical properties (e.g. the aeronet network). Retrievals of aviris 2015-2016 aerosol extinction spectra will be carried out by applying clive rodger s optimal estimation methodology. This methodology includes retrievals and error analysis in a self-consistent framework. The error analysis will indicate the limitations of what can be retrieved from aviris spectra in regards to aerosols. The ultimate product will be a suite of geographically specific aerosol spectra suitable for use in atmospheric correction for current 2015-1016 aviris-india measurements and future imaging spectrometer data taken in india. A geographically specific model of aerosols for atmospheric correction of imaging spectrometer data has never, to our knowledge been produced and this would be the first of its kind. The catalog of aerosol spectra will be freely distributed to the aviris community. More generally, we expect to achieve substantial improvements in atmospheric correction of hyperspectral data, due to our particular application of optimal estimation methodology that will utilize a priori data from multiple satellites and ground based sites.