Cloud Information Content Analysis for EPIC's Oxygen A- and B-band Channels
Abstract
The Earth Polychromatic Imaging Camera (EPIC) instrument on the Deep Space Climate Observatory (DSCOVR) will have two molecular oxygen channels: one for the well-known ``A'' band at~764 nm and one for the weaker ``B'' band at 688~nm. In both cases, a channel-integrated relative measurement of absorption is possible using an ``in-band'' channel and a nearby ``reference'' channel. Together, these four observations enable a rudimentary differential optical absorption spectroscopy (DOAS) of O2 in the characteristic retro-reflection geometry of the L1 vantage point. A priori, we thus have at best two new pieces of cloud information to access. EPIC's pixels have 10x10 km2 footprints at nadir (center of the illuminated disk), more as the viewing angle increases away from local zenith. What new information can be learned about clouds from these data on a pixel-by-pixel basis? O2 A-band observations from space have been pioneered with CNES's POLDER, ESA's SCIAMACHY, and JAXA's GOSat. NASA's OCO-2, to be launched in early 2013, will also have A-band capability. POLDER has low spectral and spatial resolutions, but offers multiple viewing directions for every pixel; SCIAMACHY has higher spectral but worse spatial resolution and just one viewing angle. GOSat has very high spectral but rather low spatial resolutions, again with the possibility of dense angular sampling, but no imaging (just one pixel at a time). OCO-2, a narrow swath imager, will have similarly high spectral resolution and reasonably high ( ∼2~km) spatial resolution. Of these four LEO missions, two are focused on CO2 DOAS, with O2 being assayed operationally only to deliver it in ppm's. POLDER and SCIAMACHY however have official cloud products based on A-band measurements. They contain, at the least, an estimate of cloud top height and, at the most, that plus an estimate of cloud pressure thickness. Cloud optical depth and effective particle size are derived from other spectral data, including continuum values near the A-band for SCIAMACHY and polarization at ``cloudbow'' angles for POLDER. This experience suggests that reasonable targets for EPIC's O2 channels are: (1) cloud top pressure, and (2) cloud pressure thickness. This assumes that cloud fraction and cloud optical depth are determined independently, that is, from the 340, 388, 443, 551~nm and O2 reference channels. We bring to bear a model for the A-/B-band signals from cloudy EPIC pixels that is simple enough to be analytically tractable. Spatially, it is based on a 3-layer scenario where absorption by a gas in the presence of scattering by particles is approximated in the cloudy layer by a diffusion-like model, and in the cloud-free layers by ballistic transport; if necessary, the surface is assumed to be a Lambertian reflector. Spectrally, the model is exercised with either filter-averaged O2 optical depths or, more correctly, by averaging the nonlinear responses over the filter function. This simplified radiative transfer (RT) model offers physical insights and, overall, substantiates quantitatively the above qualitative statements. Finally, we will show how the RT model can be generalized---and observations refined---for more spatially complex scenes (multiple and/or broken cloud layers), which will be the main vulnerability of the EPICretrieval.
- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2011
- Bibcode:
- 2011AGUFM.A43G..06D
- Keywords:
-
- 0319 ATMOSPHERIC COMPOSITION AND STRUCTURE / Cloud optics;
- 0321 ATMOSPHERIC COMPOSITION AND STRUCTURE / Cloud/radiation interaction;
- 3359 ATMOSPHERIC PROCESSES / Radiative processes;
- 3360 ATMOSPHERIC PROCESSES / Remote sensing