H layering in the top meter of Mars

We explore the capability of a method of mapping the depth distribution of a hydrogen-rich layer in the top meter of Mars from the neutron currents measured by the Mars Odyssey Neutron Spectrometer. Assuming the soil can be modeled by two layers of known composition having different hydrogen contents, simulations allow an inversion of the neutron data into knowledge of depth and hydrogen content of the lower layer. The determination of these variables is sensitive to the hypothesis of chemical composition of the soil. We quantify this contribution to the uncertainty in the method first in terms of individual chemical elements and then in terms of macroscopic absorption cross sections. To minimize this source of error, an average composition was inferred from Mars Exploration Rover data. Possible compositions having a wide range of macroscopic absorption cross sections were used to evaluate the uncertainty associated with our calculations. We finally compare our results to ice table depth estimates predicted by two published theoretical models at locations where the composition is relatively well known. The fit is excellent in the southern high latitudes but questionable in the northern high latitudes. Possible explanations of these differences include the high geographical variations of the neutron currents relative to the spatial width of the response function of the instrument and the overly simple model we, of necessity, used for surface layering.