ORIGIN OF POST GLACIAL NEGATIVE GRAVITY
ANOMALY UNDER HUDSON BAY
William A. White
Department of Geology
Mitchell Hall
University of North Carolina at Chapel Hill
Chapel Hill, NC 27599-3315
Abstract
This paper suggests that the negative free air gravity anomalies centered in Hudson Bay and the Gulf of Bothnia result mostly from removal of phanerozoic cratonic cover. It presents arguments to show that such cover was removed from the Canadian and Fennoscandian Precambrian Shields by the repetitive ice sheets that covered them in Pleistocene times.
KEYWORDS: Glaciation, Erosion, Hudson Bay, Gravity Anomaly
* * *
Negative free air gravity anomalies that neatly coincide with the areas lately (c. 10,000 years B.P.) freed from the burdens of the Laurentide and Fennoscandian Ice sheets, have commonly been ascribed to uncompleted isostatic adjustment to the loss of their former ice loads. Simons and Hager (1997) associate them with more deep seated causes, and Mitrovica (1997) observes:
Associating the entire observed peak free-air gravity anomaly over Canada with post-glacial disequilibrium is highly problematic. It has long been recognized that the relatively large gravity anomaly is inconsistent with the short relaxation time (some 2,000-3,000 years) inferred from the post-glacial uplift record from the same region. Simply put, the uplift record indicates that the post-glacial adjustment is nearly complete, whereas the gravity anomaly suggests that a significant level of disequilibrium remains.
I agree with the opinion that removal of the ice load was not the major factor in causing the negative gravity anomaly. I think the anomaly is the result of the erosion of phanerozoic cratonic cover by the erosive efforts of twenty-odd episodes of glaciation during some two million years of Pleistocene time.
Geologists have been loath to recognize the effectiveness of an ice sheets' erosive ability. Traditionally they have regarded the grooves, roches moutonees and rock-basin lakes as the result of little more than a mere skinning off of pre-glacial regolith. Nothing is more obviously false. The depth of innumerable rock basin lakes (closed, undrained, topographic depressions in bed rock) attest this assertion. Lake Superior occupies a bed rock basin 1,300 feet deep below the level of its spillway, which in turn is probably the result of an unknown further depth of ice sheet erosion. Comparison with fluvial erosion is instructive. Except in areas of no appreciable relief, rivers excavate narrow lineal valleys between wider uplands that form the major parts of their drainage areas. Ice sheets erode the entire surface they flow over and ultimately shape it into a bare rock plain of very small local relief. Witness the Canadian and Fennoscandian Shields.
Differences of mechanism are significant. Erosion of hard bare rock by rivers depends on the abrasion caused by the rock waste they bring to the site of the erosion, except in young streams, no particles bigger than sand or gravel. Such erosion is minuscule compared with that produced by a particle of rock waste embedded in the sole of an ice sheet and dragged over the bed rock surface by a mass of ice some mile or two thick.
I formerly (White 1972, 1988) offered evidence that the Laurentide and Fennoscandian Ice Sheets had been very effective agents of erosion, had shaped the major geomorphic features of the areas they formerly covered, and had removed large masses of cratonic phanerozoic strata, as well as unknown but probably smaller amounts of Precambrian basement rock. I now offer additional argument to support that idea, and suggest that the negative gravity anomalies in the Laurentide and Fennoscandian areas result mostly from the removal of such lithologic load.
Glacial geologists now accredit some twenty-odd episodes of glaciation during some two million years of Pleistocene time (Van Donk 1976, Harland 1989). I suggest that the present land surfaces of the Laurentide and Fennoscandian areas are the result of erosion accomplished during those glacial episodes, and the large negative gravity anomalies are the result of the lithologic-load removed in the process of shaping those land surfaces; by the erosive effects of the ice sheets that recurrently covered them in Pleistocene time.
At the center of the area formerly covered by the Laurentide Ice Sheet, the broad inland sea -- Hudson Bay-Foxe Basin -- occupies a large bed rock basin which is co-extensive with the maximum intensity of the negative gravity anomaly. From this basin the land surface rises radially outward, to the west, south and east to a regional divide -- the Canadian Height of Land -- which separates centripetal drainage radially inward to Hudson Bay, from centrifugal drainage radially outward to the McKenzie and St. Lawrence River systems. These two rivers occupy major parts of a great arcuate trough of transcontinental extent from Amundson Gulf in the Arctic Ocean to the Straits of Belle Isle in the Atlantic -- a distance of some 7,000 kilometers.
Along this trough, which I call "the Arc of Exhumation" (White 1972), is distributed the largest assemblage of big, deep, rock-basin lakes in the world. It is generally accepted that their basins were formed by glacial erosion. They include the following water bodies: Great Bear, Great Slave, Reindeer, Athabasca, Winnipeg, Winnepegosis, Manitoba, the five Great Lakes, the St. Lawrence Estuary, and the Straits of Belle Isle whence the trough continues in submarine form along the Labrador Sea, Davis Strait, and Baffin Bay, between the Precambrian Rock of their western shores and the submerged cratonic cover. Most of these water bodies lie in the zone of contact between Precambrian crystalline rock of the Canadian Shield and surrounding phanerozoic strata of the North American Craton. Of those that do, their inward shores, toward Hudson Bay, on the concave side of the Arc of Exhumation, are crystalline Precambrian rock, forming the edge of the Canadian Shield. Their outward shores, away from Hudson Bay, on the convex side of the Arc of Exhumation, are phanerozoic sedimentary rocks, the edge of the continuum of sediment that surfaces the North American Craton. Between their inward shores and the Canadian Height of Land lie most of 95 places where exhumed pre-Ordovician erosion surfaces sculpted from Precambrian rock have been observed by nearly as many geologists (Ambrose 1964).
Subaerial fluvial erosion could not have caused this exhumation, because it is incapable of digging a topographic basin, and in many of these instances of exhumed pre-Ordovician erosion surfaces, both exhumed surface, and remnants of the strata that formerly buried it, extend deeply below water level in the bed rock basins of lakes. Such exhumation demands an erosive agent free to erode significantly in three dimensions. Glacial erosion could serve well.
Where these remnants of basal, usually Ordovician, cratonic strata remain, it is plausible to conclude that in preglacial (pre-Pleistocene) time the entire area formerly covered by Laurentide Ice Sheets held similar stratigraphic cover commensurate in thickness with that of the non-glaciated parts of the North American Craton.
In the areas between the remnants of exhumed erosion surfaces one must assume additional erosion of underlying Precambrian basement rock. Such deeper penetration is manifest in the glaciogenic erosional land forms carved out of Precambrian rock -- ubiquitous rock-basin lakes of all sizes, great grooves, roches moutonee, etc.
If these observations be accepted, it follows that the phanerozoic stratigraphic cover of the North American Craton extended over the Canadian Shield in pre-glacial time, that it has mostly been removed by the twenty-odd successive Laurentide Ice Sheets, and that differential thicknesses of underlying Precambrian basement rock have been removed also.
Probably the best way to estimate the thickness of the cratonic strata that were removed by the ice sheets is to assume that it had the same average thickness as that which remains on the non-glaciated part of the craton.
Figure 1 shows isopachs of phanerozoic cratonic cover in Canada. Most of it lies in a circuloid zone which surrounds the basal Precambrian rock of the Canadian Shield. It passes through southeastern Ontario, Manitoba, Saskatchewan, Alberta, Northwest Territories, the Arctic Islands and submarinely, through Baffin Bay, Davis Strait and the Labrador Sea, whence it enters southeastern Quebec. Not shown are parts that are in adjacent areas of the U.S.A. The rest is more concentrated in an area near and under Hudson Bay. It may be noted that the thickness of cover shown by the isopachs increases consistently toward the outside of the circuloid zone, and toward the center of the Hudson Bay area.
Since, in the western New York and Manitoba-Saskatchewan, Alberta areas, the stratigraphic section thins by glacial erosion of the land surface rather than by lensing out of strata, I conclude that the other parts of the peripheral circuloid zone do also, and that in pre-Pleistocene time cratonic cover was continuous across the Canadian Shield from all or nearly all parts of the circuloid peripheral zone to the central Hudson Bay area, as I suggested in 1972, and as in the definitive statement made by Stott and Allen (1993):
St. Lawrence Platform lies along the southeastern margin of the Canadian Shield. . . . Its history is closely linked with the histories of the Interior and Hudson platforms, for it forms part of a more extensive cratonic cover that was at times continuous with Hudson Platform to the north and Interior Platform to the northwest.
To their statement I would add that one of those times lasted until the beginning of Pleistocene time. The presence of Ordovician, Silurian, Devonian and Cretaceous cratonic strata in both circuloid peripheral zone and Hudson Bay areas support the idea of a congenital origin.
The truncation of phanerozoic strata agrees closely with the geographic limit of the ice sheet's ability to remove it entirely, as shown by the lakes along the Arc of Exhumation. This supports the idea that those strata extended up-glacier over the Precambrian Shield in preglacial time. It is also significant that the outcrop zone of the entire thickness of the phanerozoic stratigraphic cover of the craton lies between the outer limit of glacial erosion and the exposure of Precambrian rock of the shield.
Were the outcrop zone of the phanerozoic strata the result of long enduring subaerial dissection, it should be grossly digitate in plan, with long tongues of Precambrian exposure extending down-glacier along valleys that the river had cut through the easily eroded flat-lying phanerozoic strata of the craton.
In the northern edge of the Appalachian Plateau in western New York (Figure 2) contrast between the topographic effects of fluvial and glacial erosion can be seen in the pattern of outcrop of horizontal strata on the map as it changes progressively in an up-glacier direction from the limit of glacial erosion at the south to a zone along the south shore of Lake Ontario where all remnants of fluvial topography have been excised. At the south the pattern of outcrop is tortuous with the outcrop zone of each stratum extending in a narrow band along the walls of each fluvial ravine. It gradually becomes simpler in plan with distance northward up-glacier until it becomes wide, straight, east-west extending bands, each of uniform width, on a reliefless plain of glacial scour which differs topographically from the bottom of adjacent, glacier-dug Lake Ontario only in being subaerial rather than sublacustrian. The fluvial topography has been progressively eroded away by the ice sheet until none of it remains.
This progressive change in the effects of fluvial versus glacial erosion is described here because it also shows the equally progressive removal of the preglacial, fluvially dissected topography by glacial erosion, and hence the removal of the part of the cratonic strata it had been etched out of.
This progressive removal of preglacial topography is graphically shown by the topographic model of the Elmira sheet of the 1/200,000 series of topographic maps (Figure 3). At the south, in the distance, is the fluvially dissected topography of the northern edge of the Appalachian Plateau, beyond the limit of glacial erosion. With distance northward, toward the foreground, that fluvially dissected topography has been increasingly eroded away by the Ice Sheet, until none of it remains on the flat, Lake Border Plain of Lake Ontario on the near edge of the model. With a greater distance of some 250 kilometers up-glacier from the edge of the undamaged fluvial topography the entire thickness of the cratonic cover has been eroded away by the ice sheet, and the Precambrian basement is exposed at the edge of the Canadian Shield.
The thickness of this mass of phanerozoic strata removed by the ice sheet makes it difficult to conclude that it extended no farther onto the Precambrian Shield in preglacial time than it does now, as does also the failure of any individual stratum other than the mobile salt to thin as it approaches outcrop and the Precambrian Shield.
This tendency for strata of the Phanerozoic cratonic cover to maintain their thickness to the point of exposure in the circuloid zone around the Canadian Shield is seen consistently in the four stratigraphic cross-sections across Manitoba, Saskatchewan and Alberta, presented by Gussow (1962). The cover loses thickness, not by congenital thinning of individual strata as would be true if they had been lensing out (thinning) toward the edge of the sedimentary environment in which they were formed. Thinning here occurs by simultaneous surficial erosion of the entire stratigraphic section. Each successive stratum maintains its thickness until it is beveled by erosion. The innumerable lake basins cut into bed rock assure that the erosion was glacial. Such collective thinning of the whole section can be seen in Figure 4, which is a reproduction of the pertinent eastern part of section A B B' C D, the southernmost of Gussow's (1962) four cross-sections. The other three are essentially similar.
The magnitude of glacial erosion in western New York in a narrow geographic zone, at the debilitated (thinned) equator-ward edge of the ice sheet's climatic domain demonstrates the erosive vigor of a moving ice sheet. During whatever parts of the two million years of Pleistocene time the ice was present, it has eroded a far greater mass of phanerozoic sediment than the entire time of post-Paleozoic exposure has in the adjoining nonglaciated part of the Appalachian Plateau.
How long the non-glaciated part of the Appalachian Plateau has been exposed subaerially is not known, but whatever length of time it has been exposed to subaerial denudation the whole stratigraphic section of Paleozoic strata is still extant in the center of the Permian Basin. And save for some evidence of former regional beveling by a planar erosion surface (personal communication, John M. Dennison) the entire nonglaciated part of the plateau has suffered little more than dissection in the form of narrow, steep-walled stream valleys.
That such fluvially dissected topography extended to the limit of glacial erosion is seen in the northern edge of the nonglaciated plateau immediately south of the Valley Head moraines along the New York-Pennsylvania state line (Figure 3). Yet from there northward, up-glacier, the ice sheet has left only progressively more obscure remnants of the preglacial fluvial topography until in a distance of some 95-110 kilometers it has been effaced completely to form the Lake Border Plain of Lake Ontario. In doing so, it has progressively eroded away increasingly larger parts of the phanerozoic stratigraphic section and in a longer distance of some 270 kilometers from the edge of the fluvially dissected topography the phanerozoic section, perhaps some 2,000 meters thick, has been eroded away completely to expose Precambrian rock at the southern edge of the Canadian Shield.
These and similar observations along the periphery of the Canadian Shield are strong argument for the idea that deep erosion by the Laurentide Ice Sheets has been the dominant factor in sculpting the surface they flowed across, and that such erosion removed phanerozoic cratonic cover, and in doing so has exposed the world's largest area of Precambrian rock.
Note also (White 1972) that the large, multicontinental areas of Precambrian exposure formerly covered by the Pangean Ice Sheet of Dwyka-Gondwanna time bear only post-Dwyka cover, although pre-Dwyka strata overlie the Precambrian in adjoining areas, just as pre-Pleistocene strata overlie it around the area formerly covered by the Laurentide Ice Sheets.
The difference in the means of transfer of mass is significant in the matter of isostatic adjustment. In the melting of an ice sheet the mass of the resulting liquid water is distributed throughout the world's oceans, in eustatic sea level rise. Small amounts of wind-blown dust would also be spread afar. But most of the debris produced by the ice sheets' erosive effects is distributed no farther than the distance turbidity currents can flow, debris-carrying icebergs can float before they melt away, or submarine landslides can extend. I suspect that nearly all the debris produced by the erosive effects of the Laurentide and Fennoscandian ice sheets was deposited in the abyssal basins of the North Atlantic Ocean, Baffin Bay, Davis Strait and Labrador Sea, except for that which the Mississippi River carried to the Gulf of Mexico. Thus the large positive gravity anomaly in the North Atlantic Ocean (Simons and Hager 1997) (their Figure 1-a) may result from the transfer of some 2,000 meters of phanerozoic cratonic cover from the Canadian Shield, and the debris produced in beveling the edge of surrounding cratonic cover, as well as the debris eroded in shaping the surface of Fennoscandia.
Landes (1962) presented a map showing all the salt deposits of North America (Figure 5). All but one of them are in the edge of the Paleozoic cratonic cover where it curves around the Canadian Shield forming the outer (down-glacier) valley wall of the Arc of Exhumation.
Such arcuate, linear distribution of evaporites around the edge of an area of older crystalline rock is improbable unless it and its host strata formerly extended onto the area of older rock. I suggest that the salt originated in a phanerozoic Hudson Bay sedimentary basin that was larger in area than its present remnant and was squeezed out of it laterally by halokinesis induced by the overburden of successive Laurentide Ice Sheets; a process somewhat similar to that described by Watkins (Watkins et alia, 1995; Li and Watkins 1993) in his studies of salt in the sediments of the Gulf of Mexico. (Note that a remnant of the Devonian host strata of the salt still exists in the present remnant of the Hudson Bay stratigraphic basin.)
The salt would not have been able to pierce the ice with diapirs because of its greater density but would be pressed out from under the mobile and vibratory ice load.
The ease with which salt may be deformed is seen in the desert of southern Iran where a salt diapir flows out onto the land surface on a hill and flows down the hillside in a "salt glacier" under no distorting burden other than its own mass.
I conclude that the salt was formerly present together with its Devonian host stratum in the area that is now the exposed Precambrian rock of the Canadian Shield.
As I described earlier (White 1972) all the characteristic topographic features of the Laurentide region have counterparts in Fennoscandia. The Gulf of Bothnia is the counterpart of Hudson Bay. Skagerack, Baltic Sea, Lakes Onega Ladoga, the Gulf of Dvinsk and the White Sea are the counterparts of the large water bodies of the Arc of Exhumation in North America. The Canadian Height of Land is simulated by the divide between drainage to these Fennoscandian lakes and seas and to the Gulf of Bothnia, the counterpart of Hudson Bay. A broad area of exposed Precambrian rocks lies around the Gulf of Bothnia, which is the counterpart of the Canadian Shield around Hudson Bay. As in the Laurentide area, the dimensions of the major topographic features are proportional to the diameter of the ice sheets which formed them. So also is the associated Free Air Gravity anomaly. All are smaller in Fennoscandia.
References
Ambrose, J.W. , 1964 Exhumed paleoplains of the Precambrian shield of North America. Am. Jour. Sci. 262, 817-857.
Bradsahw, B.E., and Walker, J.S., 1994. Growth fault evolution in offshore Texas, Transactions of the Gulf Coast Association of Geological Societies, v. XLIV, p. 103.
Gussow, W. C., 1962. Regional geological cross-sections of the Western Canada sedimentary cover; Alberta Society of Petroleum Geologists. Geological Society of Canada. 1962.
Harland, W.B., et alia, 1989. A geologic time scale 1989, Cambridge University Press.
Landes, K.K., 1962. Origin of salt deposits; In Symposium on Salt; Northern Ohio Geological Society, Cleveland, Ohio.
Li, Rong, Watkins, J.S., 1993. Plio-Pleistocene Growth Fault systems in West and East Cameron, and their south addition areas. Offshore Louisiana outer continental shelf, GCSSEPM Foundation, 14th Annual Research Conference: Rate of geologic processes. Dec. 5-8, 1993.
McCrossan, R.G. and Porter J.W., 1973; the geology and potential of the Canadian sedimentary basins -- A synthesis; in The Future Petroleum Provinces of Canada, R.G. McCrossan (ed.). Canadian Society of Petroleum Geologists, Memoir 1. p. 589-720.
Mitrovica, J.X. , 1997 Going halves over Hudson Bay. Nature 390, 445-447.
Simons, M. & Hager, B.H. , 1997 Localization of the gravity field and the signature of glacial rebound. Nature 390, 500-503.
Stott, D.F. and Aitken, J.D., 1993. Introduction; Chapter 1 in Sedimentary Cover of the Craton in Canada, D.F. Stott and J.D. Aitken (ed.); Geological Survey of Canada, Geology of Canada, no. 5, p. 1-7 (also Geological Society of America, The Geology of North America, v. D-1).
Van Donk, J., 1976. A record of the Atlantic Ocean for the entire Pleistocene Epoch, G.S.A. Memoir 145.
Watkins, J.S., MacRae and Simmons, G.R., 1995. Bipolar simple-shear rifting responsible for distribution of mega salt basins in Gulf of Mexico, GCSSEPM Foundation, 16th Annual Research Conference: Salt, sediment and hydrocarbon. Dec. 3-6.
White, W.A., 1972. Deep erosion by continental ice sheets. Geol. Soc. Am. Bul. 83,
1037-1056.
White, W.A., 1988 More on deep glacial erosion by continental ice sheets and their tongues of distributory ice. Quaternary Research 30, 137-150.
White, W.A., 1992. Displacement of Salt by the Laurentide Ice Sheet. Quaternary Research v. 38, p. 305-315.
