Atmospheric correction of hyperspectral ocean color sensors: application to HICO Amir Ibrahim NASA GSFC / USRA Bryan Franz, Zia Ahmad, Kirk knobelspiesse (NASA GSFC), and Bo-Cai Gao (NRL)
Remote sensing of the Atmosphere- Ocean system Solar Light interacts with the Atmosphere-Ocean system according to the following principles: - Scattering by Aerosols, hydrosols and gases - Absorption by Aerosols, hydrosols and gases - Reflection and refraction of the air-sea interface Atmosphere F 0 L t For simplicity, we assume the radiance contribution, L, of every component of the AO system to be Ocean surface additive: Satellite observations measure the total radiance at the top of atmosphere, L t L t = L aerosols + L molecules + L surface + L ocean Ocean f L aerosols f + L molecules L ocean Lsurface Ocean color enthusiasts need to measure this signal. Is that simple?.. NO
MODIS Aqua scene of the Arabian sea Natural scenes acquired from space can be complex: Absorbing aerosols Strong Sun glint Bright waters Thin clouds Dust plumes Several atmospheric correction techniques are used in NASA s ocean color operational algorithm to mitigate these problems Thin clouds Bright waters Glint
Heritage Multi-spectral spaceborne sensors MODIS Aqua example: Spectral information content: F 0 L t Ocean color bands Atmospheric f L correction aerosols and aerosol bands f + L molecules L ocean Lsurface Ocean color, AC and aerosols bands are located at Atmospheric windows (i.e., minimally impacted by absorbing gases)
Why do we need to compensate for absorbing gases in the AC? Absorbing gases including: water vapor, oxygen, ozone and nitrogen dioxide modulate the measured TOA radiance significantly within the visible spectrum. Oxygen Water vapor + ozone + oxygen features Water vapor HICO: Lake Erie, USA, Cyanobacteria blooms A correction algorithm for gases is need to removed the unwanted spectral features in ocean reflectance. Erroneous correction of gases can significantly degrade ocean color data quality and plankton type algorithms. HICO: East Coast of Australia, Trichodesmium blooms
NASA s operational multispectral AC algorithm extended to hyperspectral From L0 L t λ = L r λ + L a λ + L ra λ + t λ L f λ + T λ L g λ + t λ L w λ T g λ Hyperspectral AC & vicarious calibration process To L3 Stage 1 Stage 2 Stage 3 Stage 4 L1B calibrated top of atmosphere hyperspectral radiance (i.e. HICO, AVIRIS) Sensor specific tables for Rayleigh and aerosols Ancillary data Pressure, RH, wind speed, ozone, NO2 f gain vl t L t L r, L a, F 0 O 3, RH, WS, P, NO 2 L2GEN L2gen AC processing module T g ATREM (H2O, O3, O2, NO2, N2O, CH4, CO2, CO) 8 gases Implementation L w vl w in situ L w Level 2 products, [Chl], Rrs, Es, a, bb, water vapor, etc. Vicariously calibrated Level 2 products, [Chl], Rrs, Es, a, bb, water vapor, etc. Validation Vicarious calibration in-situ Radiance at just above sea surface (i.e. MOBY)
Hyperspectral Imager for Coastal Ocean (HICO) HICO on the Japanese Experiment Module Exposed Facility (JEM-EF) Platform International Space Station (ISS) Operation lifetime 2009-2014 Orbit repeat time/period 3 days/90 minutes Scene size (km) 50 x 200 Pixel size (m) ~100 Wavelength (nm) 353 1080 (128 bands) Spectral resolution (nm) 5.7 Spectral FWHM (nm) 10 (<= 745 nm), 20 (>745 nm) Sensor type Signal-to-noise ratio (SNR) Offner Spectrometer > 200:1 assuming 5% surface albedo Polarization sensitivity <5 %
Vicarious calibration of HICO Because of HICO s calibration problems, such as thermal instability, second order diffraction effects, and lack of an on-board calibration system, a vicarious calibration of HICO is needed to improve ocean color data quality Based on co-incident, colocated in-situ MOBY data, vicarious gains were derived At MOBY
Global in-situ to HICO validation and MODISA to HICO comparison HICO MODISA With gains No gains Chesapeake Bay, USA 2.5 hours difference
Conclusion Hyperspectral ocean color remote sensing requires a proper compensation for absorbing gases in the atmosphere. NASA s operational multispectral AC algorithm has been extended to hyperspectral to include compensation of all gases, especially water vapor and oxygen. After the vicarious calibration of HICO, derived ocean color products were improved. The hyperspectral AC algorithm for HICO is currently available through SeaDAS version 7.4 on https://oceancolor.gsfc.nasa.gov Thank you for listening!