An integrated multiangle, multispectral, and polarimetric imaging concept for aerosol remote sensing from space

Abstract

Techniques for passive remote sensing of aerosol optical and microphysical properties from space include visible, near- and shortwave-infrared imaging (e.g., from MODIS), multiangle intensity imaging (e.g., ATSR-2, AATSR, MISR), near-ultraviolet mapping (e.g., TOMS/OMI), and polarimetry (e.g., POLDER, APS). Each of these methods has unique strengths. In this paper, we present a concept for integrating these approaches into a unified sensor. Design goals include spectral coverage from the near-UV to the shortwave infrared; intensity and polarimetric imaging simultaneously at multiple view angles; global coverage within a few days; kilometer to sub-kilometer spatial resolution; and measurement of the degree of linear polarization (DOLP) for a subset of the spectral complement with an uncertainty of 0.5% or less. This high polarimetric accuracy is the most challenging aspect of the design, and is specified in order to achieve climate-quality uncertainties in optical depth, refractive index, and other microphysical properties. Based upon MISR heritage, a pushbroom multi-camera architecture is envisioned, using separate line arrays to collect imagery within each camera in the different spectral bands and in different polarization orientations. For the polarimetric data, accurate cross-calibration of the individual line arrays is essential. An electro-optic polarization "scrambler", activated periodically during calibration sequences, is proposed as a means of providing this cross-calibration. The enabling component is a rapid retardance modulator. Candidate technologies include liquid crystals, rotating waveplates, and photoelastic modulators (PEMs). The PEM, which uses a piezoelectric transducer to induce rapid time-varying stress birefringence in a glass bar, appears to be the most suitable approach. An alternative measurement approach, also making use of a PEM, involves synchronous demodulation of the oscillating signal to reconstruct the polarization state. The latter method is potentially more accurate, but requires a significantly more complex detector architecture.