Loss of Water to Space from Mars: Processes and Implications

One of the major sinks for water on Mars is the loss to space. This occurs via a complex series of processes that transport the individual atoms to the upper atmosphere, where several escape mechanisms remove them. Hydrogen and deuterium are lost primarily by Jeans escape. Non-thermal processes also remove H and D, but are only important in determining D loss at solar minimum under modern conditions. The present H loss rate is equivalent to the loss of 10^-3~pr-\micron~yr^-1 of water. The loss of oxygen is more complicated. The three main processes are indirect (or ionospheric) sputtering, solar wind pickup of O⁺, and O_2⁺ dissociative recombination. Their relative importance has varied over the history of Mars. The combined effect of the O loss processes is to remove a ~ 50~m global layer of water over the last 3.5 Gyr. Based on photochemical modeling, the loss of oxygen and hydrogen are balanced (over geological timescales) by a feedback process. During the early history of Mars, impact erosion and hydrodynamic blow-off may have removed significant water. But, it is difficult to estimate their quantitative effects. The transport of individual H, D and O atoms to the exosphere where they can escape is not completely understood. It occurs primarily via intermediate species, H₂, HD, O₂ and CO₂. The H₂ and HD are formed by photolysis of water and the odd hydrogen photochemistry. One open issue is the mechanism regulating the partitioning of D between HDO and HD (which controls the supply of D available for escape from the exosphere). The various loss processes isotopically enrich Martian water since the exospheric escape source region is depleted. Jeans escape and the transport from the lower atmosphere further fractionate hydrogen, the most useful isotopic system. Based on recent observations, the D/H fractionation factor, F ~ 0.02. Measurements of atmospheric water vapor indicate it is enriched in deuterium, with a D/H ratio ~ 5 times the terrestrial value. Since most of the water on Mars is likely to be in the form of ice, it is presumably further fractionated by ~ 0.8 due to ice/water vapor interactions. This yields an effective D/H enrichment of ~ 7 for reservoirs in equilibrium with the atmosphere. From a loss to space point of view, Martian water can be divided into three reservoirs. The first is the thin, 10 pr-\micron, atmospheric water. The second is a global exchangeable reservoir in long term isotopic equilibrium with the atmosphere. This probably encompasses the polar caps, ice in polar layered deposits and any other near surface ice or adsorbed water. The third, more speculative, reservoir is a non-exchanging reservoir (a deep sub-surface cryosphere). In addition, due to the small size of the atmospheric reservoir, difficulty in isotopically equilibrating it with the entire exchangeable reservoir, and the relatively rapid H₂ loss rate, there is also an intermediate exchangeable reservoir of ~ 4~mm. This is probably either a surface layer on the polar caps or near surface ice deposits. By assuming an initial terrestrial D/H ratio for Martian water (based on condritic meteorites) and a loss to space of ~ 50~m (based on the total O loss), the size of the exchangeable reservoir can be estimated. Two conceptual models are possible, depending on whether or not the non-exchangeable reservoir replenishes the exchangeable one. Quantitatively, the two models yield almost identical reservoir sizes, ~ 9~m (about the size of the northern polar cap). If, due to slow rate of isotopic diffusion in ice, the exchangeable reservoir actually has the same isotopic enrichment as the atmosphere, it would contain ~ 12~m of water.