ICARUS 72, 477-491 (1987) The Formation of Mercury's Smooth Plains 1 WALTER S. KIEFER AND BRUCE C. MURRAY Division of Geological and Planetary Sciences, California Institute of Technology, Pasadena, California 91125 Received October 28, 1986; revised July 2, 1987 There has been extensive debate about whether Mercury's smooth plains are volcanic features or impact ejecta deposits. We present new indirect evidence which supports a volcanic origin for two different smooth plains units. In Boreafis Pianitia, stratigraphic relations indicate at least two distinct stages of smooth plains formation. At least one of these stages must have had a volcanic origin. In the Hilly and Lineated Terrain, Petrarch and several other anomalously shallow craters apparently have been volcanically filled. Areally extensive smooth plains volcanism evidently occurred at these two widely separated areas on Mercury. These results, combined with work by other researchers on the circum-Caloris plains and the Tolstoj basin, show that smooth plains volcanism was a global process on Mercury. Present data suggest to us that the smooth and intercrater plains may represent two distinct episodes of volcanic activity on Mercury and that smooth plains volcanism may have been triggered by the Caloris impact. High-re- solution and multispectral imaging from a future Mercury spacecraft could resolve many of the present uncertainties in our understanding of plains formation on Mercury. © 1987 Academic Press, Inc. INTRODUCTION The origin of Mercury's plains is crucial to understanding both the geologic history (Murray et al. 1975, Strom et al. 1975) and thermal history (Solomon 1976, 1977) of the planet. At Mariner 10 resolution, two basic types of Mercurian plains, intercrater plains and smooth plains, can be distin- guished. Intercrater plains cover about 40% of the imaged part of Mercury and are more heavily cratered than the smooth plains. Smooth plains cover about 15-20% of the imaged area of Mercury (Trask and Guest 1975). The formation of the intercrater plains was discussed by Trask and Guest (1975), Malin (1976), Strom (1977), and Leake (1981). Only the smooth plains will be considered here. Because of lighting and resolution condi- California Institute of Technology Division of Geo- logical and Planetary Sciences contribution No. 4417. 477 tions, diagnostic volcanic landforms, such as domes and flow fronts, are difficult to identify on Mariner 10 images of smooth plains (Schultz 1977, Malin 1978). Only a few such features have been postulated so far. These include several craters located on or near smooth plains which may have been endogenetically altered (Schultz 1977), several irregular, rimless pits which may be volcanic collapse structures (Strom et al. 1975; Schultz 1977), and a possible pyroclastic deposit (Schultz 1977). The ab- sence of recognizable volcanic landforms does not disprove a volcanic origin, how- ever, because under Mariner 10 lighting and resolution conditions it is also difficult to identify known volcanic landforms on the Moon (Malin 1978). In the absence of defi- nitive evidence for volcanism, two alterna- tive hypotheses developed concerning the origin of the smooth plains. One group (Murray et al. 1975, Strom et al. 1975, Trask and Strom 1976) argued that the 0019-1035/87 $3.00 Copyright © 1987 by Academic Press, Inc. All rights of reproduction in any form reserved.478 KIEFER AND MURRAY smooth plains are volcanic features analo- gous to the lunar maria; the second group (Wilhelms 1976, Oberbeck et al. 1977) con- tended that the smooth plains are ejecta deposits, analogous to the lunar Cayley plains. Because direct evidence for volcanism on Mercury is lacking, advocates of vol- canism attempted to show that Mercury's smooth plains could not be produced as ejecta deposits, and thus must be volcanic in the absence of a viable alternative. For example, Strom et al. (1975) noted the presence of partially filled ghost craters on the floors of several basins which contain smooth plains. Such craters must have formed after the basin but prior to plains formation. This proves that these plains are not ejecta deposits from the basin-forming impact. However, it leaves open the possi- bility that the plains are ejecta deposits from younger, distant basins. As the largest and youngest known basin on Mercury, Caloris is an obvious potential source basin for ejecta deposits. Strom et al. (1975) suggested several reasons why Caloris ejecta cannot explain the observed smooth plains. Strom et al. were unable to detect the presence of ancient basins in the region surrounding Caloris. They assumed that other basins must have formed in this region and are now buried. From this, they estimated an average thickness of 10 km and a total volume of 5 × 10 7 km 3 for the circum-Caloris smooth plains. They argued that this volume of plains material is too large to be derived solely as Caloris ejecta. However, more detailed geologic mapping of this region has revealed evidence for numerous basin rings (Spudis and Strobell 1984, Spudis and Guest 1987). This discov- ery removes the rationale for the Strom et al. thickness estimate and hence invalidates their volume estimate. A study of partially filled craters on the smooth plains indicates an average plains thickness of less than 1 km, comparable to the thickness of the lunar mare (DeHon 1979). DeHon's thick- ness estimate assumes that all partially filled craters formed on the preplains sur- face, so the amount of crater filling is a direct measure of the plains thickness. If some of the ghost craters studied by DeHon actually formed on a preexisting smooth plains surface, then the average plains thickness is somewhat larger than DeHon estimated. Nevertheless, there is currently no evidence to support the Strom et al. estimate of the average smooth plains thickness. Arguments based on the total volume of circum-Caloris smooth plains should therefore be viewed with skep- ticism. Strom et al. also considered the spatial distribution of smooth plains