ABSTRACT A large portion of the Belt-Purcell Supergroup is well exposed in the vicinity of Glacier and Waterton Lakes national parks of northwestern Montana, USA, and southwestern Alberta, Canada. These strata were deposited in the northeastern part of the Mesoproterozoic Belt Basin. The dramatic rate of subsidence combined with dominantly fine-grained sediment influx produced thick units of broadly uniform lithology, which constitute the spectacular and unusually colorful mountain scenery of this region. Seemingly fairly simple at first glance, in detail these rocks exhibit a great deal of facies heterogeneity and a number of unusual attributes. This has resulted in contrasting and controversial interpretations of sedimentary features, depositional dynamics, sedimentary environments, and consequently the overall understanding of the entire basin. The Belt Basin reveals itself to be a unique setting in many respects, but ideas stemming from these rocks have implications for other strata, not just those of pre-Cambrian age, but for the entire Phanerozoic as well. The Belt Supergroup is therefore a particularly stimulating field-trip destination that challenges textbook interpretations.

The ca. 1.460 Ga Revett Formation is a gray and purple quartzite lithosome in northwestern Montana, and it interfingers eastward into red argillite of the Grinnell Formation in Glacier National Park. The Revett Formation was analyzed in northwestern Montana by identifying sedimentary structures in stratigraphic sections and by interpreting flow processes of the structures using the standard flow regime model (e.g., Simons et al., 1965). The sedimentary structures and thicknesses of the event beds were then organized into eight sediment types (lithofacies) that were grouped into three sediment complexes: the playa complex, the antidune complex, and the sheet sand complex. The arrangements of the sediment types and complexes within the stratigraphic framework of the lower informal Revett member indicated the configurations of the depositional environments in space, and the vertical configurations of the sediment types revealed the depositional history of the lower Revett member. The lower Revett member lithosome interfingers eastward into the red argillite of the Grinnell Formation lithosome, and has eight through-going descriptive, stacked, lithic units, called lithostromes. Lithostromes 2, 4, 6, and 8 (from the bottom up) are composed of the sheet sand complex and extend into playa complexes of the Grinnell Formation. They were deposited by sandy sheetfloods that flowed at grade and terminated as the water sank into the sand substrate. Between lithostromes 2, 4, 6, and 8 are lithostromes marked by playa lakes of the playa complex that spread from the east across western Montana during humid periods. They were overlain by sheetfloods of the antidune complex that built eastward over the playa complex as the playa lakes retreated with increasing aridity. The antidune complex was overlain by the sheet sand complex of a vast sand plain deposited by sheetfloods from the southwest that flowed at grade level across western Montana during arid periods. The sheetflood deposits of the Revett Formation were mostly deposited by the upper-flow regime element of the established fluvial facies model.

Crinkle cracks are sand-filled cracks up to 5 mm wide in plan view that pinch at their ends. In cross section, they are canted and crinkled. They cut mudstone beds that underlie hummocky cross-laminated sandstone lenses. They are here described from the Piegan Group, Proterozoic Belt Supergroup, but they are widespread in Proterozoic and Phanerozoic marine and lacustrine rocks. However, they represent a distinctive, descriptive style of mudcracks, not attributed to inferred syneresis processes, although they have been commonly attributed to syneresis. In plan view, crinkle cracks closely resemble cracks formed where oscillatory waves striking viscous mud banks are transformed into fluid solitary-like waves that open surface cracks on their trailing limbs and close the cracks on their leading limbs as they pass through the viscous mud. Crinkle cracks preserved in rocks are hypothetically attributed to oscillatory waves moving sand over viscous mud. The oscillatory waves are transformed into solitary-like waves as they pass down into the mud, forming the cracks. The surface sand falls down into the cracks, preserving them. With burial, the water escapes, and the viscous mud compacts, crinkling the sand-filled cracks.

The Chamberlain Formation, one of the lower members of the early Mesoproterozoic Belt Supergroup, has previously yielded low-diversity assemblages of microfossils but the reported fossils were of limited utility for inferring paleoenvironmental conditions. Here, we describe substantially more diverse microfossil assemblages from drill core of the Chamberlain Formation obtained from the Black Butte mine locality near White Sulphur Springs, Montana. The Chamberlain Formation biota contains abundant Valeria , Leiosphaeridia , Synsphaeridium , and Lineaforma , with lesser amounts of Satka , Symplassosphaeridium , and Coniunctiophycus. The assemblages partially overlap with, but are distinct from, microfossils recently reported from the Greyson Formation, another unit from the Helena embayment of the Belt Supergroup. Since the overlapping taxa exhibit similar states of preservation but dissimilar relative abundances, we interpret the assemblages as reflective of distinct paleoenvironmental conditions of the sampled sections of the Chamberlain and Greyson Formations. The Chamberlain Formation assemblages are most comparable to microfossil groupings reported from the Bylot Supergroup of Canada and the Roper Group of Australia from sediments from very shallow-water (supratidal to lower shoreface) marine environments. This comparison corroborates previous hypotheses on the basis of sedimentological data that the lower Chamberlain Formation sediments were formed in a lagoonal or mud-flat environment. By contrast, the Greyson Formation assemblages are most comparable to microfossil groupings associated with sediments from shallow-shelf marine environments. The fidelity of comparisons among the 1.2 Ga Bylot Supergroup, 1.49 Ga Roper Group, and 1.45 Ga Belt Supergroup assemblages indicates that the groups of microorganisms that produced these assemblages, and their associations with the paleoenvironments that they inhabited, may have been characteristic of the littoral marine biosphere throughout much of the Mesoproterozoic.

New regional mapping documents that a thick quartzite sequence in the Lemhi subbasin of the Belt-Purcell basin lies near the top of the Mesoproterozoic stratigraphic column, and that two finer-grained units have been miscorrelated. This observation requires reassessment of the subbasin's stratigraphy, which we present here. Determination of the relationships between the stratigraphic units of the Lemhi Range and Salmon River and Beaverhead Mountains and better-known Belt Supergroup units to the north has been hampered by miscorrelation of this upper quartzite sequence with older strata, and by miscorrelation of the type Apple Creek Formation with a similar but stratigraphically lower unit. The base of the upper quartzite sequence includes the Swauger and Lawson Creek Formations, which are the highest units previously identified in the Lemhi subbasin. This sequence continues upward through quartzite units described here that underlie or comprise lateral equivalents of the type Apple Creek Formation in the Lemhi Range. The spatial distribution of these quartzite units extends the Lemhi subbasin farther east and north in Montana and northwest in Idaho. The complete stratigraphy reflects the stratigraphic separation of the two “Apple Creeks” and expands the type Apple Creek Formation to accommodate the quartzite units into the regional Mesoproterozoic stratigraphy. Our proposed correlation of the thick upper quartzite sequence with the Bonner Formation and higher units of the Missoula Group in the main part of the Belt basin requires that subsidence of the Lemhi subbasin was significantly faster than that of the main part of the Belt basin during deposition of the upper Missoula Group. Therefore, the two parts of the Belt basin were distinct tectonically, although they shared common sediment sources.

A recent 1:24,000 scale mapping project within the northern Beaverhead Mountains along the Idaho-Montana border has resulted in a reinterpretation of both the Mesoproterozoic stratigraphy and the regional structural framework. A 15-km-thick stratigraphic section of the Mesoproterozoic Lemhi subbasin was initially deformed by northeast-southwest shortening into giant northwest-striking, northeast-verging folds, probably during Cretaceous Sevier orogenesis. These initial folds were then dissected by a system of subparallel and anastomosing, oblique-slip reverse, thrust, and normal faults that generally strike northwest, but that exhibit east-west–oriented lineations, suggesting components of strike-slip displacement. Contractional faulting appears to have been followed by Eocene to Miocene extensional faulting, with many normal faults following the preexisting fabrics. Extension opened Tertiary basins along some of these faults, including the Salmon Basin along the southwestern side of the Beaverhead Range. Subparallel faults in the surrounding region appear to have a similar complex history, and all appear to be part of a major northwest-striking Cretaceous fold-and-thrust belt that was later dissected by Tertiary extension. Although the faults of the Beaverhead Mountains are significant and long-lived, they are not terrane-bounding structures separating the Belt and Lemhi sedimentary sequences. Instead, Lemhi strata extend across the range and northward to Missoula, where they grade into correlative Missoula Group strata.

We used laser ablation–inductively coupled plasma–mass spectrometry to determine the U-Pb ages for 1206 detrital zircons from 15 samples of the Lemhi subbasin, upper Belt Supergroup, in southwest Montana and east-central Idaho. We recognize two main detrital-zircon provenance groups. The first is found in the Swauger and overlying formations. It contains a unimodal 1740–1710 Ma zircon population that we infer was derived from the “Big White” arc, an accretionary magmatic arc to the south of the Belt Basin, with an estimated volume of 1.26 million km 3 —a huge feature on a global scale. The ɛ Hf(i) values for magmatic 1740–1710 Ma zircons from the Lawson Creek Formation are +8–0, suggesting that they were derived from more juvenile melts than most other Lemhi subbasin strata, which have values as evolved as −7 and may have been derived from an arc built on Proterozoic or Archean crust in the Mojave Province. Since paleocurrents in cross-bedded sandstones indicate northward flow, the proximate source terrane for this sand was to the south. The second provenance group is that of the Missoula Group (and Cambrian strata recycled from the Missoula Group), with significant numbers of 1780–1750 Ma grains and more than 15% Archean grains. This provenance group is thought to represent mixing of Yavapai Province, Mojave Province, and Archean Wyoming Province sources. Both of these provenance groups differ from the basal Belt Prichard Formation, and strata of the Trampas and Yankee Joe Basins of Arizona and New Mexico, which contain a major population of 1.61–1.50 Ga non–North American grains. The 12 youngest grains from the several Swauger Formation samples suggest the formation is younger than 1429 Ma. The three youngest grains from Apple Creek Formation diamictite suggest the rock is younger than 1390 Ma. This makes the Apple Creek diamictite the youngest part of Belt Supergroup strata south of the Canadian border. Though the Big White magmatic arc was produced before 1.7 Ga, the sediment may have been recycled several times before being deposited as locally feldspathic sandstone in the Lemhi subbasin depositional site 300 m.y. later. Because the detrital-zircon provenance does not change from Idaho east to Montana, our data do not support the existence of a major Great Divide megashear separating the Lemhi subbasin from the Belt Basin. In southwest Montana, unfossiliferous sandstones of Cambrian age contain the same detrital-zircon assemblages as the Swauger Formation and Missoula Group, suggesting reworking of a local Belt Supergroup source.

The redox state of the mid-Proterozoic oceans, lakes, and atmospheres is still debated, but it is vital for understanding the emergence and rise of macroscopic organisms and eukaryotes. The Appekunny Formation, Belt Supergroup, Montana, contains some of these early macrofossils dated between 1.47 Ga and 1.40 Ga and provides a well-preserved record of paleoenvironmental conditions. We analyzed the iron chemistry and mineralogy in samples from Glacier National Park, Montana, by pairing bulk rock magnetic techniques with textural techniques, including light microscopy, scanning electron microscopy, and synchrotron-based X-ray absorption spectroscopy. Field observations of the Appekunny Formation combined with mineralogical information allowed revised correlations of stratigraphic members across the park. However, late diagenetic and/or metasomatic fluids affected primary iron phases, as evidenced by prevalent postdepositional phases including base-metal sulfides. On the west side of the park, pyrrhotite and chlorite rims formed during burial metamorphism in at least two recrystallization events. These complex postdepositional transformations could affect bulk proxies for paleoredox. By pairing bulk and textural techniques, we show primary records of redox chemistry were preserved in early diagenetic and often recrystallized framboidal pyrite, submicron magnetite grains interpreted to be detrital in origin, and red-bed laminae interpreted to record primary detrital oxides. Based on these observations, we hypothesize that the shallow waters of the mid-Proterozoic Belt Basin were similar to those in modern marine and lacustrine waters: fully oxygenated, with detrital reactive iron fluxes that mineralized pyrite during organic diagenesis in suboxic, anoxic, and sulfidic conditions in sedimentary pore waters.

The Neihart Quartzite and LaHood Formation are the lowermost units exposed in the Helena embayment, which forms the eastern and southeastern margins of the Belt Basin. Ages of detrital zircons from the Neihart Quartzite (quartz arenite) and a range of lithologies in the LaHood Formation (conglomerates to arkoses to siltstones) show that these units do not share a common provenance. The dominant provenance is Paleoarchean for the LaHood Formation and Paleoproterozoic for the Neihart Quartzite. Provenance is further constrained by the geochemistry and U-Pb ages of zircons from cobbles from the classic LaHood conglomerate in Jefferson Canyon (Tobacco Root Mountains), ages of Paleoproterozoic crystalline basement in the Beaverhead-Tendoy Mountains (1.8–2.45 Ga), and elemental and Sm-Nd isotopic data for select samples of both sedimentary rocks and crystalline basement within the basin. These data show a pronounced lack of detritus from abundant, proximal Neoarchean (2.7–2.9 Ga) and Paleoproterozoic (1.9–2.5 Ga) crystalline basement exposed in Laramide uplifts and the soles of Sevier-style thrust faults within and near the basin. Analyses of detrital mineral assemblages in the Lower Belt Supergroup units clearly indicate that the finer-grained portions of the LaHood Formation were not locally derived, based on abundant white mica in sections overlying tonalite-trondhjemite-granodiorite (TTG) basement and lack of amphibole in units overlying hornblende tonalites. Significant fractionation also exists between sand- and cobble-size components in conglomerate of the LaHood Formation in terms of elemental abundances, isotopic compositions, and the U-Pb ages of zircons. Stratigraphically, the differences in the ages of the youngest zircons in all LaHood Formation samples and the Neihart Quartzite (1.71 Ga, Neihart; 1.78 Ga, LaHood) do not refute any proposed stratigraphic correlations. Nonetheless, age spectra of detrital zircons from the Neihart Quartzite, all LaHood lithologies, and previously published data for the Newland Formation show distinctions of provenance and an apparent lack of interaction among the sediment-supply systems of these three formations. This contrast suggests that distinct, likely fault-bounded, sedimentologically restricted subbasins characterized the initial stages of development of the eastern Belt Basin along the Perry line (southeastern margin of the Helena embayment), in the manner of a modern, but partially submerged, Basin and Range topography. The time of development of this topography is not clear, but it may have been related to the collapse phase of the Great Falls orogeny at ca. 1.7 Ga for the Helena embayment. The primary, north-south–trending Belt Basin also developed subsequent to the Great Falls orogeny along the western paleomargin of the newly amalgamated Wyoming–Medicine Hat–Hearne craton.