Geological Survey Professional Paper 754—C The Portage Lake Volcanics (Middle Keweenawan) on Isle Royale, Michigan

The Portage Lake Volcanics is the upper part of the total Keweenawan volcanic sequence, one of the world's major accumulations of plateau or flood basalt flows. This sequence, together with the associated Keweenawan continental red clastic rocks, is probably more than 40,000 feet thick locally. These materials fill a trough or series of elongate basins nearly 100 miles wide and more than 1,000 miles long, coincident in part with the midcontinent and mid-Michigan high gravity and magnetic anomalies (Bayley and Muehlberger, 1968). The structure has been interpreted as a major rift system of continental proportions. (See, for example, Hinze and others, 1971; King and Zietz, 1971; Lyons, 1970; Pettijohn, 1970.)

The source of the lavas appears to have been along the axis of the trough or rift system, as directional flow data on the Keweenaw Peninsula indicate flow from the Lake Superior basin toward the peninsula, and the Portage Lake Volcanics has been interpreted as wedging out completely within 15-20 miles south of the present line of outcrop of the sequence on the peninsula (White, 1960). On the north shore of Lake Superior in Ontario, Tanton (1931, p. 64) noted that in a number of localities flowage wrinkles on the tops of lava flows in the Osler Group indicated flowage toward the north or northeast prior to consolidation. He concluded that "all observations indicate that the lava came from a southerly source and it seems probable that the lava was poured out along fissures in the site of Lake Superior." Although the Osler Group is probably stratigraphically below the Portage Lake Volcanics on Isle Royale, the two sequences appear to have many similarities, and no reason now indicates that the lavas came from different sources. Definitive directional flow data are lacking on Isle Royale, but stratigraphic considerations previously mentioned suggest that the Isle Royale vicinity was not far from the northwest margin of the depositional basin at the time the Portage Lake Volcanics was erupted. Assuming this location, there is almost no alternative but that great volumes of lava were erupted from fissures near the axis of the trough and spread laterally and rapidly toward both margins of the trough, finally ponding and cooling in place to form vast sheets covering literally thousands of square miles.

White (1960, p. 372-373) summed up the mechanism of basalt flooding as follows:

The principal consequence of the great volume of lava floods is that on a relatively flat surface a flow may come approximately to rest while its interior, at least, is still molten, giving the top time to approach a hydrostatic level. The near-level surface of such a flow provides the floor for a succeeding flow, making the upper and lower surfaces of successive flows nearly parallel over large areas. As envisaged here, the formation of a basalt flood has its closest observed counterpart in the initial burst of lava from vents on shields. The great volume of lava and the gentle slope keep the flow thick, molten, and exceedingly wide, and do not permit the top to become anchored or tunnels to form. Motion stops primarily because the lava is ponded, either against an uphill slope or behind the dam of its own frozen margin, After eruption ceased, there would, of course, be some movement of lava from central to marginal parts of the flow, and some local advances of the front itself as long as the leveling process continued, but these movements would be minor compared with those of the initial flooding. The crystallization of a large part of the interiors of flows would thus take place under essentially static conditions.

White arrived at his interpretation largely on the basis of field evidence, speculating that the great volume of lava alone, rather than an abnormally high fluidity, favored the formation of flood basalts instead of the construction of shield volcanoes of the Hawaiian type, which are made up of rock types that do not differ greatly.

Shaw and Swanson (1970a, b) recently calculated some eruption and flow rates for flood basalts on the basis of certain physical parameters for the Yakima Basalt flows of the Columbia River Plateau and measured viscosity data for Hawaiian basalt. They (1970a) concluded the following:

Fluid dynamical calculations suggest average flow velocities in the turbulent regime of 10 to 20 km per hour for flows of about 10 m or more in thickness assuming average effective viscosities of approximately 500 poises. Such velocities indicate that heat loss during flow was mainly by radiation, and that lavas 200 km from their source cooled an average of about 50°C during flow. Turbulent flow implies homogeneity of flow units and little development of surface crust during movement, and the small temperature depression during flow implies that quenched lavas would have low crystal contents if they were initially at or above their liquidus. Both implications are consistent with field studies.

They further concluded that with only minor and altogether realistic changes in their assumptions, chiefly somewhat larger source fissures, surface flow rates approaching 50 km per hr can be calculated. The field evidence suggesting rapid eruption and flowage, with ponding and cooling in place, is not inconsistent with theoretical flow mechanics.

Data from the Keweenaw Peninsula and Isle Royale indicate that the clastic debris in the sedimentary rocks interbedded with the lava flows of the Portage Lake Volcanics and in the Copper Harbor Conglomerate and other Keweenawan formations above the lava sequence was transported into the basin from marginal source terranes (White, 1952, 1960; White and Wright, 1960; Hamblin and Horner, 1961; Wolff and Huber, 1973). The evidence for basinward flow of streams, contrasted with evidence that the lavas flowed toward the margins of the basin, shows that there were at times reversals of the prevailing slope over large areas, and the picture of a basin subsiding as it was filled seems established beyond any reasonable doubt. According to White's analysis (1960), the lava was horizontal or sloped gently toward the margins of the basin as long as filling kept pace with downwarping. When extrusion was interrupted, however, downwarping produced basinward slopes over small to large areas, permitting sedimentary debris to be swept into the basin during those periods of volcanic quiescence. Finally, with the gradual cessation of volcanic activity, continued subsidence permitted the accumulation of the Copper Harbor Conglomerate and younger deposits to form a thick sedimentary sequence in the basin.

The gross synclinal form of the Keweenawan basin resulted from subsidence coincident with filling of the basin rather than later folding. However, attitudes of strata near the margins of the basin, as on the Keweenaw Peninsula and Isle Royale, were subsequently steepened by movement on major thrust faults, the Keweenaw fault and the Isle Royale fault, thereby accentuating the synclinal structure.