Reanalyzing MAHLI Images from the Bagnold Dune Field: Lessons in Uncertainty and Production of the Largest Known Dataset of Martian Grains

Abstract:

Aeolian bedforms are ubiquitous on the surface of Mars and have been investigated for several decades. Like their terrestrial counterparts, understanding the grain size distributions (GSDs) of Martian bedforms can provide essential context for understanding the processes that underpin their emergence and development. The Mars Hand Lens Imager (MAHLI), used during Curiosity’s Bagnold Dune Field Campaign at Gale Crater, has provided high-resolution images from which grain size metrics can be acquired for individual Martian bedforms. MAHLI images have been central to recent interpretations of Martian bedform properties and dynamics. Here we take a deeper dive into the interpretation of image-based GSDs from Martian bedforms to highlight new insight. Using MAHLI images of bedforms and a novel digitizing approach that allows for some quantification of uncertainty and volumetric estimates, we have generated the largest known dataset of Martian grains. Using our dataset and reanalysis of previously published datasets, we argue that these data are not equivalent to those used to derive terrestrial GSDs due to the limitations presented by image-based grain size analysis. These limitations include: (i) the miniscule sample sizes involved, (ii) distortion of GSDs in terms of grain counts versus grain volume, (iii) limitations from surface-only grain size measurement and the anisotropy of GSDs in bedforms, and (iv) uncertainties of image-based approaches. We present estimates of volumetric grain distributions to demonstrate that the cumulative volume of sediment digitized on Mars from several datasets amounts to a fraction of a teaspoon and that several MAHLI images previously characterized as monodisperse or unimodal may instead be poorly sorted or multimodal. Our results suggest that although it is tempting to use image-based GSDs based on grain counts to support Martian aeolian process interpretations, the limitations of the data are such that these interpretations cannot be made with nearly the level of confidence applied to terrestrial features. In the face of this uncertainty, a culture of ‘Martian exceptionalism’ should not be acceptable and all Mars grain size data and process interpretations based on these data should be treated speculatively until large-scale sample return or analysis is feasible.