April 2004 LIP of the Month

Lunar mare basalt volcanism: Formation of Large Igneous Provinces (LIPs) on a one-plate planet.

James W. Head
Department of Geological Sciences, Brown University,
Providence RI 02912 USA
james_head@brown.edu

Large igneous provinces are common on the Moon, Mars and Venus and their characteristics offer a potentially important perspective for interpreting LIPs on the Earth. In this overview on the Moon drawn from Head and Coffin (1997), the focus is on the characteristics of the Earth's Moon (see also Hiesinger and Head, 2004). The Moon is about one-quarter the diameter of the Earth and is of lower density, has not retained an atmosphere, is characterized by vertical tectonics of an unsegmented lithosphere (not lateral plate tectonics), and presently has a very thick lithosphere. Most of its geological surface activity took place in the first half of solar system history (Wilhelms, 1987). Basaltic volcanic deposits occur on the Moon in the form of the lunar maria, which cover about 17% of the surface, primarily on the near side (figure 1). The total area of the lunar maria (~6.3 x 106 km2) is considerably larger than typical terrestrial LIPs occurrences but only slightly larger than the area of the Ontong Java large igneous province (Head and Coffin, 1997) (figure 2). The lunar maria were emplaced over a period of about 3 billion years, largely in the first half of solar system history, but the total volume was relatively small, about 1 x 107 km³. This value for the total planet is comparable to the volume of the terrestrial Deccan flood basalt deposits alone, but considerably less than the total present volume of the terrestrial oceanic crust, about 0.17 x 1010. The average lunar global volcanic flux was low, about 10-2 km³/a, even at peak periods of mare emplacement (in the Imbrian Period, 3.8-3.2 Ga). This average global flux is comparable to the present local output rates for such individual terrestrial volcanoes as Kilauea or Vesuvius. Output rates for individual eruptions on the Moon could be extremely high, however. Several individual eruptions associated with sinuous rilles may have emplaced more than 103 km³ of lava in about a year, a single event that would represent the equivalent of about 70,000 years of the average flux!

Volcanic features that might indicate large-volume eruptions and large accumulations include the individual maria themselves, extensive flow fronts, some stretching for distances of over 1200 km (Schaber, 1973), volcanic complexes that might signal the location of hot spots, and sinuous rilles, which have been attributed to high effusion rate eruptions involving thermal erosion of the substrate (Hulme, 1973; Carr, 1974). Interestingly, no large shield volcanoes, such as those seen on the Earth (e.g., Hawaii), Mars (e.g., Olympus Mons), or Venus (e.g., Sapas Mons), are observed on the Moon (Head and Wilson, 1991); large caldera-like features are also extremely rare.

The lunar maria themselves are of diverse sizes and shapes and individual occurrences might be thought of as equivalent to some terrestrial large igneous provinces, particularly those deposits that tend to be concentrated within large impact basins of various states of preservation (Head, 1976). Indeed, Alt et al. (1988) have proposed that large terrestrial lava plateaus that form within plates with no apparent tectonic cause are the terrestrial equivalents of the lunar maria. In their model, an impact crater on Earth large enough to cause pressure-release melting would be quickly flooded to form a lava lake (equivalent to the lunar maria) and these events, in turn, would initiate hot spots which would develop into persistent low pressure cells within the mantle (Alt et al., 1988). Does the lunar record support this model? Although early theories suggested a causal relationship between lunar impact basin formation and basaltic mare filling, the results of the Apollo and Luna exploration programs, and models of basin formation and evolution (Solomon et al., 1982; Bratt et al., 1985), showed that generation of basalts was unlikely and that there was a temporal separation between impact basin formation and mare basalt filling.

In the case of the 900 km diameter Orientale impact basin, vast quantities of substrate were impact-melted by the basin-forming event to produce a smooth sheet of plains lining the basin interior and floor and estimated to have a volume of ~200,000 km³ (Head, 1974). This relatively smooth plains unit could be easily confused with a large volcanic province. The unit, however, has a compositional affinity to the non-mare target rocks (Head et al., 1993) and is distinctly different in composition and time of emplacement from the adjacent basaltic maria deposits, which span an interval of several hundred million years (Greeley et al., 1993).

In most other mare basins, the vast majority of the exposed volcanic plains were emplaced over an extended period of several hundred million years following the impact event (Basaltic Volcanism Study Project, 1981; Hiesinger et al., 2000, 2003). There is no evidence for the production of basin-sized lunar basaltic "lava lakes which crystallized from the surface down" (Alt et al., 1988). The stratigraphic record of lunar maria infilling shows well the long and sequential development, and the difference in age between the basin-forming event and its basaltic lava filling. Localization of the maria in the basins was apparently due to passive variations in crustal thickness and ponding in topographic lows (Head and Wilson, 1992).

Although the equivalency of an impact origin of a basin and its fill on the Moon and large igneous provinces on Earth proposed by Alt et al. [1988] is not supported by evidence from the Moon, the possibility that impact events on Earth could initiate volcanism should not be completely discarded. The small size of the Moon (and correspondingly very different pressure gradient), its thicker crust and variable lithosphere thickness, could all inhibit melting relative to a comparable event on Earth. Convincing arguments have been put forth to indicate that impact-initiated volcanism (in contrast to direct impact melting) was not a factor in the large (~200 km diameter) Sudbury basin formed in continental crust on Earth (Grieve et al., 1991). Similar-sized impacts into thin crust and lithosphere typical of a young oceanic floor setting could conceivably produce pressure-release melting and associated volcanism (e.g. Jones et al., 2002). The relatively small size and shallow depths of excavation of craters typically formed during the time of emplacement of most well-documented large igneous provinces (e.g., the last 250 Ma; Coffin and Eldholm, 1994) makes impact cratering a much less likely candidate for their formation and evolution than large-scale rifting and deep-source plume volcanism, although exceptions certainly can occur. Recently, Ivanov and Melosh (2003) have outlined the issues that must be addressed to make a convincing case for impact-initiated volcanism on Earth.

Other large volcanic accumulations on the Moon include the extensive lava flow fronts of Mare Imbrium (figure 3) which were emplaced at least a billion years after the formation of the impact basin and occur in three phases which extend 1200, 600, and 400 km from the southwestern edge of the basin into the basin interior. These three flow units have a total volume of >4 x 104 km³ and very high effusion rates are implied by their lengths and volumes; effusion rates and flow volumes are comparable to some of those observed in the Columbia River flood basalts (Schaber, 1973). The fact that these units are some of the youngest on the Moon suggests that other more degraded flows filling the earlier lunar maria may also have been emplaced in a comparable mode. Examination of isolated mare basalt ponds in the highlands fringing the continuous maria has shown that typical volumes are in the range 100-1200 km³, values similar to those of terrestrial basaltic flood eruption units (Yingst and Head, 1994; 1995). Thus, the individual eruptions that make up the maria may be equivalent to units within flood basalt and large igneous provinces on Earth, but the eruption frequency seems to be much less; the lunar maria are emplaced over many hundreds of millions of years, rather than a few million years as is apparently the case in the terrestrial examples.

Another unusual characteristic of lunar mare deposits relative to those on Earth are the occurrence of sinuous rilles, which are meandering channels occurring primarily in the lunar maria, and preferentially along the edges of the maria. They range in widths up to about 3 km and in length from a few km to more than 300 km. Sinuous rilles are generally an order of magnitude larger and often much more highly sinuous than terrestrial lava channels. Many of the characteristics of lunar sinuous rilles unexplained by simple lava channel/tube models can be accounted for by thermal erosion (Hulme, 1973; Carr, 1974). The characteristics of large sinuous rilles (their length, width, depth, and the nature of their source regions) provide important information on eruption conditions. For a 50 km long rille in the Marius Hills, Hulme (1973) calculated an effusion rate of 4 x 104 m³/s, an eruption duration of about one year, and a total volume of about 1200 km³. The characteristics of source depressions of sinuous rilles led Wilson and Head (1980) and Head and Wilson (1980) to describe independent evidence for extremely high effusion rate eruptions of long duration. On the basis of these studies, key factors in the formation of sinuous rilles by thermal erosion are: 1) turbulent flow, requiring high effusion rates and aided by low yield strength; and 2) sustained flow (implying very long eruptions and thus very high eruption volumes) to cause the continued downcutting to the observed rille depths. Thus, eruptions that caused many of the large sinuous rilles found on the Moon were apparently characterized by rapid effusion of low yield strength lavas for prolonged periods producing flows of extremely high volumes (in the range 300-1200 km³), comparable to those in terrestrial flood basalt provinces. In contrast, typical eruption volumes for shield-related flows on Earth are much less than a km³ (Peterson and Moore, 1987), with the largest historic lava flow (Laki) being about 12 km³.

Several areas show unusual concentrations of volcanic features on the Moon (Guest, 1971; Whitford-Stark and Head, 1977). Two of the most significant of these are the Marius Hills area (35,000 km2), which displays 20 sinuous rilles and over one hundred domes and cones, and the Aristarchus Plateau/Rima Prinz region (40,000 km2), which is dominated by 36 sinuous rilles. The high concentration of sinuous rilles suggests that these complexes are the site of multiple high-effusion rate, high-volume eruptions, and that these centers may be the surface manifestation of hot spots and possible analogs to terrestrial large igneous provinces. The thick crust and lithosphere characteristic of the Moon (and thus the greater depths of sources) may help to make these types of features less recognizable, and more analogous to continental volcanic provinces. In addition, deposits on the Moon are much more widely dispersed distally from their sources.

In summary, the lunar maria are comparable in scale to some terrestrial large igneous provinces but appear to have been emplaced over much longer periods of time (e.g., 108 to 109 years rather than 106 to 107 years). Many individual eruptions, however, appear to be similar in volume and eruption rates to those in flood basalt provinces. There is little evidence for shallow magma reservoirs and repeated small-volume eruptions that would build up large shield volcanoes. These summary characteristics seem to call for large batches of magma erupted over short periods of time from relatively deep sources, but separated in time by significant intervals. How can these characteristics be accounted for in terms of the nature of the source regions and the modes of emplacement?

The basaltic maria are superposed on the ancient globally continuous and thick low-density anorthositic highland crust derived primarily from global-scale melting associated with planetary accretion. This low density highlands crust provided a density barrier (Solomon, 1975) to mantle plumes and basaltic melts ascending from the mantle. Rising diapirs and magma bodies may thus have tended to collect at the base of the 60-80 km thick crust. Following sufficient overpressurization of source regions, individual dikes propagated toward the surface. Thus, the thick highlands crust may have created a deep zone of neutral buoyancy for rising magma that could only be overcome by overpressurization events propagating dikes to the surface.

Whether there was intrusion or eruption might have been determined by variations in overpressurization and crustal thickness. Low levels of overpressurization resulted in intrusion into the lower crust, forming dikes which cooled and solidified. Dikes that were characterized by sufficient overpressurization to approach the surface could have several fates. Intrusion close enough to the surface to produce a distinctive near-surface stress field often resulted in the production of linear graben-like features along the strike of the dike and small associated effusions and eruptions. In the case of Rima Parry 5, small spatter cones are aligned along the central part of the linear graben (Head and Wilson, 1994a). Overpressurization events that are large enough to propagate dikes to the surface are predicted to have very large volumes (Head and Wilson, 1992). Indeed, predicted volumes are comparable to those associated with many observed lava flows, such as the flows extending hundreds of kilometers into Mare Imbrium (Schaber, 1973), and those associated with sinuous rilles.

The relationship between magma source size and highland crustal thickness may have been such that the frequency of dikes propagating to the near-surface and surface to produce eruptions was relatively low, but there is controversy about the actual processes and factors governing ascent and eruption (e.g., Wieczorek et al., 2001). A low frequency of dike emplacement events would mean that most dikes had sufficient time to cool before the next dike was emplaced. Thus, emplacement of a plexus of dikes at a sufficiently high frequency to create a shallow reservoir may have been very difficult on the Moon. The lack of Hawaii-like shield volcanoes on the Moon and the paucity of caldera-like features is thus attributed to the difficulty in producing shallow reservoirs which would then result in the emplacement of many individual flows to form an edifice and associated calderas (Head and Wilson, 1991). In addition, the same lack of multiple, continuous dike emplacement events of such magnitude to reach the surface over short periods of time may have meant that the lunar maria large basaltic provinces tended to be produced from relatively large eruption events over very long periods of time, in contrast to terrestrial large igneous provinces.

The lunar situation may be analogous in many ways to basaltic magma bodies interacting with terrestrial continental crust. Here, zones of neutral buoyancy (Glazner and Ussler, 1988) cause buoyantly rising basaltic magma bodies to stall in the crust. Overpressurization events can cause the same features seen on the Moon, as exemplified by many of basaltic volcanic fields in the western United States (e.g., Crumpler et al., 1994), and indeed large-scale flood basalts can be emplaced that are comparable to the large lunar flows (Tolan et al., 1989). The low melting temperature of the continental crust relative to that of the more refractory lunar anorthositic crust means that stalled basaltic magma bodies in continental crust may cause associated and large-scale melting, resulting in a complexity not known to occur on the Moon. The continental crust and the lunar highlands illustrate the role of large-scale density barriers to the creation of significant shallow basaltic reservoirs, such as those observed beneath the ocean floor and in large edifices such as Hawaii. Complex shallow reservoirs do exist in continental crust, however, where local conditions of melt generation and, unlike the Moon, sustained supply rates, exist (as in continent margin subduction zones and central hot spot traces or rifting environments).

In summary, the Moon may provide examples of LIPs deposits formed by infrequent but very voluminous eruptions, where as much as 70,000 times the annual volcanic flux emerges in a single eruption in the space of a single year!

References