Optical Depth Retrievals from Shadows in HiRISE Images

The spatial resolution of the images that the HiRISE camera is taking of the Martian surface is unprecedented and offers new ways to estimate the optical depth τ of the abundant aerosols in the Martian atmosphere. While resolved shadows are rare in images of Mars at spatial resolutions of tens or more meters per pixel, HiRISE’s images actually often resolve the abundant smaller shadows that are cast by boulders and cliffs in the rims of craters. We use a HiRISE image of the Opportunity rover site to estimate optical depths from shadows and compared our estimates with the ground truth that the rover measured by looking at the sun. Our results suggest that retrievals of the optical depth from shadows can yield an accuracy of better than 10%. 1) Introduction: The optical depth of aerosols in the Martian atmosphere frequently is substantial [1]. The scattering by airborne dust and other aerosols diminishes contrast and changes color in images taken from orbit. The optical depth must (at least roughly) be known before one can attempt to correct for these atmospheric effects. However, in the visual it is not at all trivial to quantify optical depths accurately from orbiter images. Various standard methods that are often used in Earth remote sensing do not (yet) work for Mars. E.g., the difference between measured TOA albedos (Top Of Atmosphere albedos) and known surface albedos can yield estimates of the optical depth. See for instance the documentation on MISR’s aerosol retrieval algorithms that is available at: http://eospso.gsfc.nasa.gov/eos_homepage/for_scientists/atb d/docs/MISR/atbd-misr-09.pdf. Orbiter observations of Earth can thus often be used to estimate optical depths. For orbiter observations of Mars however, this is much more problematic since ground truth albedos generally are not well known. Moreover, various standard optical depth retrieval algorithms need elaborate aerosol models as an input [2]. For Martian aerosols, such models do exist for the common airborne dust [3], but condensation of small amounts of ice onto these dust grains and icy high altitude haze layers commonly change the aerosol properties. The High Resolution Stereo Camera (HRSC) of the European orbiter Mars Express improved the situation considerably since it is often possible to retrieve optical depths from its images with the so called ‘stereo method’ [1]. However, the method is not very useful for the images of other orbital cameras like MOC, THEMIS, and HiRISE because these generally do not provide usable stereo information. Radiation that is scattered in the atmosphere is by far the most important source for the brightness that shadows display in space observations. With increasing optical depth in the atmosphere the scattering increases and the shadows grow in brightness. Thus, the brightness of shadows contains information on this optical depth. Sect. 3 offers theory concerning the so called ‘shadow method’ which we use to estimate optical depths from the brightness of shadows. Sect. 4 offers details on the HiRISE image we used. Since the Opportunity rover provided ground truth (τ = 0.32; [4]) we were able to use the image to explore, to some extent, the possibilities, limitations, and accuracy of the shadow method. We offer results and some discussion in Sect. 5 and conclusions in Sect. 6. Until the images from HiRISE started coming in, the shadow-method was of limited use because images of Mars rarely show shadows at spatial resolutions above about ten meters per pixel. The DTMs (Digital Terrain Models) from HRSC have made it clear that at such resolutions almost all slopes on Mars are quite shallow. Only when the Sun gets below 30° above the horizon do shadows begin to appear, but then the shadow method (as we developed it) quickly becomes inaccurate because it uses a plane parallel approximation which then loses its validity. However, imaged at considerably higher spatial resolution the Martian surface often shows spots that are in true shadow; e.g., behind cliffs in the rims of small craters. Many HiRISE images are so sharp that such shadows are well resolved, and it seems worthwhile to further develop and test the shadow method for HiRISE imagery. Once the optical depth is known aerosol models [3] can be used to correct for atmospheric effects. 2) Instrument: The High Resolution Imaging Science Experiment (HiRISE) is flying onboard the Mars Reconnaissance Orbiter (MRO) mission. It can image on a variety of spatial scales down to about 30 cm per pixel. For images taken in the red filter, such as the one used in this work, the signal to noise ratio is about 200:1. See http://hirise.lpl.arizona.edu/ and [5] for more information on the camera. 3) Theory: Let τ be the optical depth of the atmosphere, B i j ( , ) be the upward radiation from the surface before extinction by the atmosphere in pixel i j , of an image I i j ( , ) . Let A i j ( , ) be the contribuSeventh International Conference on Mars 3226.pdf