Follow this link to skip to the main content

LogoMISR abstract

Comstock, R.L., and Bills, B.G., (2004). Topography of the Flattest Surface on Earth: Using ICESAT, GPS, and MISR to Measure Salt Surface Morphology on Salar de Uyuni, Bolivia. Eos Trans. AGU, 85(47), Fall Meet. Suppl. 2004, Abstract # C33B-0345


Salt flats are aptly named: they are composed largely of salt, and are maintained as nearly equipotential surfaces via frequent flooding. The salar de Uyuni, on the Altiplano in southwestern Bolivia, is the largest salt flat on Earth, with an area of 9,800 km$^{2}$. Except for a few bedrock islands, it has less than 40 cm of relief. The upper-most salt unit averages 5 m thick and contains 50 km$^{3}$ of nearly pure halite. It includes most of the salt that was in solution in paleolake Minchin, which attained a maximum area of 60,000 km$^{2}$ and a maximum depth of 150 m, roughly 15 kyr ago. Despite ~10 m of differential isostatic rebound since deposition, the salar surface has been actively maintained as an extraordinarily flat and smooth surface by annual flooding during the rainy season. We have used the strong optical absorption properties of water in the visible band to map spatial variations in water depth during a time when the salar was flooded. As water depth increases, the initially pure white surface appears both darker and bluer. We utilized MISR images taken during the interval from April to November 2001. The red and infra-red bands (672 and 867 nm wavelength) were most useful since the water depth is small and the absorption at those wavelengths is quite strong. Nadir pointed MISR images have 275 m spatial resolution. Optical absorption of water has been used previously for bathymetric mapping, but usually in much deeper water. Our application was particularly well suited to obtaining high accuracy results, since the water clarity is high, the substrate reflectivity is high and quite uniform, and we can measure it directly when the surface water is gone. To aid in our evaluation of water depth variations over the salar surface, we utilized two sources of direct topographic measurements: several ICESAT altimetry tracks cross the area, and a 40 x 50 km GPS grid was surveyed to calibrate ICESAT. A difficulty in using these data types is that both give salt surface elevations relative to the ellipsoid, whereas the water surface will, in the absence of wind or tidal disturbances, follow an equipotential surface. Geoid height is not known to the required accuracy of a few cm in the central Andes. As a result, before comparing optical absorption from MISR to salt surface topography from GPS or ICESAT, we removed the longest wavelengths from both. Our model agrees quite well with the GPS grid and ICESAT tracks (5 cm RMS misfit) and shows a very flat and level, but not completely featureless salt surface. The most prominent topographic feature is a peripheral moat, or depression near the edge of the salar, which is most pronounced in locations near to sources of fresh water input during the rainy season. Another prominent feature is a series of wave-like ridges, with 20-30 cm amplitude, and 4-5 km wavelength. They are mainly found near the western edge of the salar, and have crests and troughs oriented mainly N-S. The process of formation of these features is still enigmatic, but seems related to wave action in the water during wet episodes.


[Full text not yet available]


Updated: 14-Jan-2005