Terrestrial Analog Studies of Mawrth Vallis, Mars: the Painted Desert

Introduction: Mawrth Vallis, Mars, exhibits unique stratigraphy characterized by distinct phyllosili-cate units across an expansive region [1,2]. A potential analog for this site is the Painted Desert in Arizona, where striking profiles of clay and iron oxide units color the landscape [3]. Both sites contain multiple phyllosilicate units in similar stratigraphy across wide regions. We have acquired near infrared hyperspectral AVIRIS images over the Painted Desert at resolutions of 3-20 meters per pixel. The objectives of this study are: 1) to compare the hyperspectral images of the Painted Desert in order to gain insights into the Mawrth Vallis region of Mars, and 2) to analyze the hyperspectral images of the Painted Desert in the context of planetary exploration, by attempting geological interpretation from remotely sensed-data, and subsequent ground-truthing to validate our interpretations [4]. Painted Desert analog site: The Painted Desert is the name of the desert region in northern Arizona, characterized by badland morphology and colorful layering [2]. It encompasses an area of approximat-mately 400 km 2 , and extends from the southeastern rim of the Grand Canyon to the Petrified Forest. It is a part of the Triassic Chinle Formation and consists of stratified layers of phyllosilicates of varying compositions [3]. Many of these layers were formed by the deposi-tion of volcanic ashes in lacustrine and fluvial environments. The process of deposition and alteration of the ashes resulted in the formation of bentonite, and also in the burial of organics in an anoxic environment. AVIRIS image processing: Two different hyper-spectral image lines in the wavelength range of 0.36 – 2.5 µm were obtained at spatial resolutions of up to 3 meters per pixel using AVIRIS. Correction for atmospheric opacities was performed using commercially available radiative transfer software (ACORN) specifically written for use with AVIRIS data. Correction processing discards data in the 1.3-1.4 and 1.78-2.0 µm regions due to saturation by atmospheric water bands in these spectral regions. Spectral endmember extraction was performed using standard ENVI spectral analysis tools, including minimum noise fraction analysis , pixel purity index, and spectral angle mapper target finder. These tools have been developed in parallel with AVIRIS and are optimized for such datasets. Results: Spectral and morphological analysis of the AVIRIS spectral cubes presents an astounding minera