Introduction

Salt beds of the Lucas Formation, Detroit River Group (Devonian), have been mined by solution by The Dow Chemical Company at Midland, Michigan (Fig. 1), since the 1940's. Descriptions of cores and other subsurface records saved over the years were used in a study begun in 1969 of individual salt beds at the plant site. Correlation of evaporite cycles solved an immediate problem of salt-bed correlation, which proved to be related to previously unrecognized local facies change. An informal numbering system developed for cycles in the Lucas Formation at the Midland plant site has been used in The Dow Chemical Company's drilling projects as far westward as Ludington, Michigan, and eastward to near Sarnia, Ontario (Fig. 1). Numbered cycles, and the terms "upper salts," "middle salts," and lower salts" used in this paper are strictly informal.

Basis for the Study

Study of individual salt beds was undertaken to improve the degree of confidence in correlation based on lithology alone. Descriptions of the old cores from closely spaced wells at Midland, Michigan, revealed remarkably continuous vertical successions of evaporite lithofacies in core after core, evidence that periods of increasing brine concentration were followed by periods of increasing dilution in patterns repeated many times (Fig. 2). Early in the study evidence became apparent that periods of increasing salinity or dilution could be inferred and correlated. The greatest relative salinity attained in a cycle--termed an "evaporite maximum" or "maximum"--is used herein as it was used by Andrichuk (1954); conversely the terms "evaporite minimum" and "minimum" imply the greatest relat ve freshening between cycles.

Correlation of depositional cycles was based on the assumption that any given cycle should have been the result of many factors acting on a single body of water. Sea-level

End_Page 73-------------------------

fluctuation, degree of restriction, evaporation rate, volume of brine in the system, influx rate, rate of refluxion of brine currents out of the basin, temperature, and other factors, should have influenced cycles in a basin-wide manner. The parent brine must have vacillated gradually from highly saline to nearly normal (Fig. 3). The aggregate effect of all factors should have driven the system toward greater salinities early in the development of any individual cycle, followed by the reversed sequence during the dilution phase, late in a cycle. Ideally, this would result in definite lithologic patterns in an ascending sequence: fossiliferous limestone, nonfossiliferous limestone, dolomite, anhydrite, low-bromine halite, high-bromine halite, potash and bitterns, high-bromine halite, l w-bromine halite, anhydrite, dolomite, nonfossiliferous limestone, and fossiliferous limestone. The cycle could be "reversed" at any point and the more soluble rock types would be progressively rare. The normal sequence would build to an evaporite maximum, indicated by the most soluble mineral in any given cycle. Most cycles of the Lucas Formation reached evaporite maxima in the anhydrite range, and none went beyond the stage of low-bromine halite.

At some time, the mass effect of factors influencing an evaporite cycle would change in the direction of increasing dilution. Gradual dilution would result in a complete and uninterrupted cycle, shown by a reversed vertical sequence of rock types of decreasing solubility, as described above. Andrichuk's (1954, p. 77) idealized cycle

Fig. 1. Location of study area, and extent of salt in the Lucas Formation (after Gardner, 1974).

End_Page 74-------------------------

Fig. 2. Description of core, Dow 6 Salt well, showing cyclicity of evaporites.

End_Page 75-------------------------

from southern Saskatchewan shows a reversed sequence of salt, anhydrite, brown, dense and saccharoidal dolomite, and dense or fossiliferous fragmented limestone, completing the "offshore" cycle with an evaporite minimum approaching normal sea water. Gardner's (1974, p. 113) example of a cycle masimum of the Lucas Formation is overlain by a reversed sequence of anhydrite beneath dolomite.

The reversed sequences were noted by Briggs (1959, p. 52).

The Detroit River rocks throughout the Michigan Basin exhibit many evaporite cycles that reflect progressive changes in deposition from carbonate minerals to anhydrite, to halite, and the reverse. In the rocks not containing salt, the alternation is between carbonate rock and anhydrite. Almost invariably the cycle is complete...

Briggs (1959, p. 52) considered the evidence that cycles in the Lucas Formation in Michigan were "complete" and that sequences were reversed as an indication "that the basin did not completely desiccate during or following times of salt deposition, but rather the basinal waters became somewhat less saline when the balance favored greater influx over evaporation." Evaporite minerals in the upper half of the cycle are products of the freshening or dilution phase and they are precipitated from brines less concentrated than immediately earlier brines. This is true whether the section includes dolomite on anhydrite, anhydrite on halite, or halite of decreasing bromine content on potassium salts, as seen in the Silurian A-1 Salt at Midland, Michigan (Matthews and Egleson, 1974). Mechanics o origin of such evaporite deposits--the "inverse sequence" or "recessive" of Richter-Bernburg (1972, p. 35)--are not understood clearly and frequently are not discussed; yet beds of inverse sequence account for almost half the volume of evaporites in the Lucas Formation.

Although reversal of sequence is normal for evaporite cycles in the Lucas Formation, it is not the case in many other evaporites where the dilution phase left little or no record, and interrupted-cycles or "half-cycles" resulted. Flooding after desiccation would tend to build "half-cycles," as would supratidal deposition. In describing a typical sabkha cycle of beds of anhydrite over dolomite in the Rainbow Anhydrite of Alberta, Bebout and Maiklem (1973, p. 324) commented that "In some cases the sequence will not be complete vertically and may not occur laterally over the entire buildup. This lack of completeness may be a result of either nondeposition or erosion, the latter being a common feature of the subaerial supratidal environment."

Fig. 3. Diagrammatic illustration of different lithologic sequences from one evaporite cycle. Correlations of maximum and minima define parallel time lines (broad dashed lines) that cross wedge-shaped lithologic boundaries (narrow dashed lines). At the locality on the left, greater salinity was maintained throughout.

End_Page 76-------------------------

Methods

Because The Dow Chemical Company's solution mining is located near the southern limit of the halite facies in the Lucas Formation, an effort was made to obtain records from wells more nearly central in the depositional basin, where more cycles that contain salt were expected, and where salt beds would be thicker and more likely to be noted accurately. Oil- and gas-test records in Gladwin, Arenac, Missaukee, Isabella, Bay, Midland, and Saginaw Counties (Fig. 1) were searched for wells that penetrated the salt (hundreds), that were described by detailed sample logs (a few), sonic or caliper-type logs (a few), and cores of salt above the "sour zone" (none). The Sun Oil Company 1 Mills Estate, Sec. 4, T 18N, R2W, Gladwin County, about 27 mi (43 km) northwest of Midland, Michigan, was sele ted as the best record of a basinal well (Fig. 4A). This well became the informal "type-section" from which all evaporite cycles described in this paper were defined.

Cycles were marked on strip logs opposite midpoints in the rock section representing maximum salinity, considered to be the "maximum" of each evaporite cycle. Evidence of greatest relative freshening or decrease in relative salinity between cycles was considered to mark the position of the "minimum" ending a cycle and beginning a new one. The problem of where the minimum reversal actually occurred was ignored; for example, whether an anhydrite between two salt beds terminated an old cycle "the covering anhydrite" of Fodemski (1972), or began a new cycle, was avoided by assuming that the reversal was at the halfway point, and that deposition continued during the salinity minimum.

Correlation of evaporite cycles and lithology was started with the Sun 1 Mills, and a numbering system of the sedimentary cycles was carried throughout the cored wells. Wells with caliper logs or similar logs were included in the cycle-numbering system. From these records, and some drilling-time logs, salt beds could be located reasonably accurately, but cycles without salt could not be identified. Other types of electrical and radioactivity logs were not suited to detailed study of the evaporite cycles. Wireline logs have since been used to identify salt-bearing sections as distant as Ludington, Michigan, 120 mi (193 km) west of Midland (Fig. 4B). Salts penetrated in wells drilled by The Dow Chemical Company at Ludington are believed to be correlated with those at Midland; in any cas , local correlations are excellent and although cores are not available, the concept of correlation of depositional cycles has helped in positive identification and mapping of individual salt beds in this area near the western edge of the halite facies.

In the Midland area, where core descriptions were available, correlation of cycles generally is superior to correlation based on lithology alone. In this region, where evaporite strata undergo marked lateral facies change, correlation is good, in spite of the fact that a given cycle may not be recognizable at all localities. Cycles shown by sequences of anhydrite-halite-anhydrite-halite may be "lost" if traced into a single bed of anhydrite or halite in directions normal to facies strike. Likewise cycles characterized by repetition of dolomite and anhydrite are not identifiable where shoreward facies change results in a single stratum of dolomite. Andrichuk (1954, p. 79) also cited instances in which "several discrete cycles...coalesce" to give the appearance of "one cycle."

Stratigraphy

Detroit River Group

In the Michigan basin, the Detroit River Group is composed of the Sylvania Sandstone, the Amherstburg Formation (the "Black lime"), the Lucas Formation, and the Anderdon Limestone, in ascending order. Several studies of the lower evaporite sections that contain "sour zone" oil reservoirs--and the productive Richfield zones somewhat lower in the section--were published by geologists of Sun Oil Company, who had developed an informal, detailed terminology (Fugate, 1968; Wirth, 1968; Sutton, 1968). Their work demonstrates clearly the numerous carbonate-sulfate-carbonate

End_Page 77-------------------------

cycles below the stratigraphic position of interest for solution mining at Midland, Michigan. The general cyclic nature of evaporites of the Detroit River Group has long been recognized (Sloss, 1953; Briggs, 1959; Ehman, 1964; Gardner, 1974).

Stratigraphic terminology for the Lucas Formation was advanced by Ehman (1964). He used 119 control points in Michigan to divide the Detroit River Group above the Black lime by using mechanical log markers. He mapped the aggregate salt within four units with no attempt to identify specific salt beds other than the "Massive salt," which he called the "Big salt" or "G-H member," and which thins consistently all directions from a maximal 90 ft (27 m) in Roscommon and Clare Counties. Ehman recognized that this salt grades into anhydrite outside the central basin area, and that it is dolomitic farther from the center.

Fig. 4. A. Location map showing stratigraphic cross sections, A-A^prime and B-B^prime. The Dow Chemical Company's Midland plant is located at the site of the Dow 6 Salt well (6S). B. Chart showing the most soluble rock type in cyclic strata identified in wells located from Gladwin County to Ludington, Michigan.

End_Page 78-------------------------

In a more recent regional study, Gardner (1974) dropped the term "Anderdon" and divided the Lucas Formation into three units: the "Richfield Member" and "Freer Sandstone," the "Iutzi Member" (or "Massive anhydrite"), and the "Horner Member." The Horner member contains at least eight salt units (Gardner, 1974, p. 32); actually of the 35 or more evaporite cycles in the Sun 1 Mills well, 19 contain halite. Briggs (1959, p. 52) cited a well in Missaukee County with at least 23 evaporite cycles. In general, distances between wells in the central basin and the lack of released data on cores prevented detailed studies of the salt section above the sour zone. In this study are included data from 20 wells in Midland County (Fig. 1) in which the salt-bearing sections were cored.

Terms used by Sun Oil Company, Ehman (1964), and Gardner (1974) are shown in Figure 5, in relation to the informal cycle-numbering system used in this paper.

Richfield Zone and Massive Anhydrite

The basal unit of the Lucas Formation is the Richfield zone, which contains numerous evaporite cycles indicated by anhydrite in dolomite; this unit can be identified in these oil fields: East Norwich (Fugate, 1968), Enterprise (Wirth, 1968), Headquarters (Sutton, 1968) and Beaver Creek (Gardner, 1974). In the Sun 1 Mills well (Sec. 4, T21S, R24W), a predominantly anhydrite section at the base of the Richfield, known informally as the "Big anhydrite," was designated cycle 1 (Fig. 5). The Richfield zone is capped by the Massive anhydrite, which in the Mills well was assigned to cycles 2 and 3, although it certainly contains evidence of several depositional cycles. These cycles are discontinuous a short distance north of Midland, and can not be identified clearly at Midland.

Lower Salts or "Sour Zone"

The Sour zone "is generally considered to include those beds between the base of the massive salt and the top of the massive anhydrite." (Fugate, 1968, p. 73). At Midland, this section of rock was assigned to cycles 4 through 11 (Fig. 5). Three of these cycles (7, 8, and 11) contain salt in the central part of the basin. Salt in cycle 7 is not in any wells drilled by The Dow Chemical Company, but the cycle is indicated by anhydrite and can be traced to Dow 92 Brine well (Fig. 5). A salt bed in cycle 8 was described in the core from Dow 27 Pressure well north of Midland (Figs. 4, 5). Cycle 9 is present in the central part of the basin as anhydrite, showing salt only at Dow 92 Brine well (Fig. 5). The salt in cycle 11 is thin in the Sun 1 Mills well and does not extend into Midland Coun y; however, an anhydrite in the stratigraphic position of cycle 11 can be traced to the Dow 31 Pressure well south of the Midland plant (Fig. 5). Other cycles below the Massive salt are stratigraphically equivalent to anhydrite in the Sun 1 Mills; these cycles may be related to salt facies northward.

Massive Salt

The Massive salt is a relatively thick, clear marker used to separate the lower salts from the middle salts. In this study, the term "Massive salt" is used, following the practice of Sun Oil Company, but Ehman (1964) and Gardner (1974) used the name "Big salt." The Massive salt is the result of four and perhaps five depositional cycles, but only two major cycles were evident at this stratigraphic position in the Sun 1 Mills well (Fig. 5). This salt may be entirely halite north of the Mills well, but in the Mills it shows an anhydritic minimum, the result of slight dilution of the parent brine. Four depositional cycles are evident within the Massive salt in the Dow 6 Salt well (Fig. 6) and at Bay City, Michigan, a fifth cycle is suspected, but is not illustrated (Fig. 7).

A few carbonate strata were of importance in this study. In most instances the rocks are dolomite, but below cycle 12 is a limestone that at Midland is evidence that halite of cycle 12 has been penetrated (Fig. 6). Beds of salt in cycles 12 and 13 are clean, but they are separated by an anhydritic minimum that is dolomitic at some localities. Salt in cycle 13 is thinner and more contaminated to the south of Midland. Salt beds of cycles 12 and 13 form a unit that is the deepest consistently mineable salt at Midland. These beds are in every well at Midland, but change southward to a predominantly

End_Page 79-------------------------

Fig. 5. Cross section A-A^prime, north-south (Fig. 4A), showing correlation of evaporite cycles from Gladwin County to Saginaw County.

End_Page 80-------------------------

Fig. 5. Continued.

End_Page 81-------------------------

anhydrite facies in the Dow 31 Pressure well, where the salt of cycle 13 is only 0.1 ft (3 cm) thick.

Middle Salts

Cycle 14 is anhydrite throughout much of the area but contains halite in a few wells at Midland (Fig. 7). Salt in cycle 15 is not in all wells at Midland (Fig. 8). However, the maximum of cycle 15 is apparent in all wells, and the time line drawn through the maxima of cycle 15 is the datum for several stratigraphic cross sections (Figs. 6, 7, 9).

Cycle 16 is shown by thin anhydrite in the Sun 1 Mills, but was not identified southward (Fig. 5). Cycle 17 includes salt in the Mills well and in some wells at Midland. It is known as one of the "variable" salts (Fig. 8, 10, 11). Salt is in cycle 18 the Mills well, but it could not be traced southward. Cycle 19 is represented by salt in the Mills well; the salt can be identified only as far southward as the Amerada 1 Letts well in southern Gladwin County (Fig. 5).

A bed of salt representative of cycle 20 extends southward beyond the Dow 31 Pressure well (Fig. 5). At Midland, this salt shows evidence of being within two cycles (20a and 20b, Dow 6 Salt well, Fig. 6).

Cycle 21 contains salt as far as the Dow 41 Pressure well, and cycle 22 contains salt that terminates south of the Dow 41 Pressure. However, salt of cycle 22 shows an anhydrite minimum at Midland, and is illustration that two cycles become apparent only near the zero edge of a salt facies (Fig. 5). Salt beds of cycles 20, 21, and 22 are separated by thin anhydrite minima and the cycles thus can be traced south across Midland County. However, cycles 23, 24, and 25 cannot be identified with confidence south of the Sun 1 Mills (Fig. 5).

Fig. 6. Detailed portion of cross section A-A^prime (Fig. 4A). Massive salt contains evaporite cycles 12a, 12b, 13a, and 13b. Cycle 14 contains halite throughout only part of the total distance covered.

End_Page 82-------------------------

Fig. 7. Cross section B-B^prime, east-west (Fig. 4A), showing correlations of evaporite cycles from Midland County to Bay County. Note that cycle 17 is apparent in both Dow 28P and Dow 1E, but the unit does not contain salt in the Dow 28P well. Correlation based strictly on lithology might correlate erroneously the salt in cycle 15, Dow 28P, with salt in cycle 17, Dow 1E.

End_Page 83-------------------------

A dolomite marker between cycles 26 and 27 was used to differentiate between Upper salt beds and Middle salt beds at Midland. Its position is clear only as far south as the Dow 28 Pressure well (Fig. 5).

Upper Salts

The Sun 1 Mills well shows evidence of nine cycles above the dolomite marker; five well-developed salt beds overlain by four anhydrite beds are designated as the nine maxima of cycles 27 through 35. Four of these salt beds extend as far south as the Amerada 1 Letts well, but the upper cycles are lost because of facies change and poor data (Fig. 5).

The amount of accurate (core) control is sparse in the section above cycle 22; consequently, salt beds of cycles 23, 24, and 26 were not defined as confidently as in cycles 12 through 22. The bentonite marker discovered by Baltrusaitis (Sloss, 1969; Baltrusaitis, 1974) is within the Upper salts (Fig. 5). It can be located on local gamma-ray logs, but only two wells were cored shallow enough to cut the zone, and apparently the bentonite was recovered in one core. Its position relative to the numbered cycles can be estimated only in the Dow 6 Salt well, where the core was described by C. K. Lucas (formerly of the Dow Chemical Company, written commun.) as including 6 in. of micaceous rock with very shaly partings.

Fig. 8. Locations of wells cored through cycle 15. In wells shown by ruled patterns, salt of cycle 15 is absent.

End_Page 84-------------------------

Fig. 9. Cross section D-D^prime, east-west, showing a detailed description of closely spaced cores within part of the solution-mining area at Midland, Michigan. "Variable salts" in cycles 14 and 17 are apparent. Cycles generally are complete, as in well 13S, but a few cycles are interrupted. For example, in well 24S the salt of cycle 15 is overlain by dolomite, whereas normally it is overlain by anhydrite.

End_Page 85-------------------------

Fig. 10. Portion of cross section B-B^prime (Fig. 4A), with added detail of a caliper log of a nearby well. Note that the "variable" salt in cycle 17 does not appear in Dow 2E well.

End_Page 86-------------------------

Facies Changes

Continuity of Deposition

Evidence of gradual change in depositional environments is apparent in cores as an unbroken vertical succession of evaporite lithofacies; in the Dow 6 Salt well, a core 627 ft (191 m) long shows only one apparent "break" in the regular succession of rock types (Fig. 2). In seven of the cores, there are no apparent interruptions, no lithofacies are "out of place" and no clays or other clastics are distinguishable in the column. Descriptions of cores seem to document an autochthonous succession of evaporite-mineral facies, similar to the "offshore" evaporite cycle Andrichuk (1954, p. 77) described as from "an area effectively removed from the influence of source...of clastic material."

A few interruptions are apparent in about two-thirds of the core descriptions; these involve a missing rock type--for example, halite on dolomite without an intervening anhydrite. Some of these interruptions may be results of secondary-mineral alterations, but they are considered here to be minor, local "disconformities," representing a brief span of time when a small part of the seafloor was exposed to air or to less-dense surface waters, if the basin contained layered brines (Sloss, 1969). The small areas showing interruptions of a given cycle are a few hundreds of meters in length and width, suggesting that the conditions that led to re-solution or nondeposition were limited to small areas (Fig. 8).

Fig. 11. Map showing edges of halite facies in cycles 14, 15, and 17, the "variable" salts, at the Midland plant site. Salt is present throughout this area in cycles 12, 13, 20, 21, and 22.

End_Page 87-------------------------

The observation that these minor "disconformities" are relatively rare and that an unbroken vertical succession of facies is normal suggests that whatever the mechanics of deposition, gradual change was prevalent in that deposition. Depth of water necessary to insure a smooth transition from one rock type to the next, whether adjacent and covering rocks were more soluble or less soluble, may have ranged from a few inches to depths approximately equal to thickness of the strata within individual cycles--no more than about 60 ft (18 m). Regardless of the manner of deposition, depositional continuity did exist in the Lucas Formation and lithologic correlations are generally straightforward from well to well, particularly along facies strike. In addition, evaporite cycles can be correlate throughout the area even in most of the cycles involving marked change in facies.

I submit that correlations of evaporite maxima define time lines, and that the correlative minima among cycles also are synchronous within about 300 sq mi (777 sq km) of Midland County, and probably across the Michigan basin. Similar conclusions were reported by Andrichuk (1954, p. 77) who stated that from southern Saskatchewan to southern Alberta, "The evaporite maxima are probably continuous."

Facies Change Within the Massive Salt

The Massive salt is an easily illustrated example of lateral extension of cycles with essentially parallel time lines that pass through nonparallel lithologic boundaries. For example, the Massive salt in the Sun 1 Mills well, Gladwin County (Fig. 4A, 5), shows evidence of two major cycles. Within 39 mi (62.8 km) from the center of the basin, the two cycles undergo the following changes (Fig. 12):

1. From midpoint to midpoint of the carbonate minima above and below the Massive salt, the two-cycle unit thins toward the margins of the basin from 92 to 79 ft (28 to 24 m) within 39 mi (62 km). This is the rate of 0.33 ft/mi (6 cm/km).

2. From the top of the basal carbonate to the base of the capping carbonate the abbreviated two-cycle unit thins from 82 to 66 ft (25 to 20 m), within 39 mi (62 km), the rate of 0.4 ft/mi (8 cm/km).

Fig. 12. Correlation of evaporite cycles 12 and 13, composing the massive salt and its facies, to illustrate the nearly foot-for-foot synchronous deposition of halite and sulfate rocks. (Locations of wells shown in Fig. 4A, cross section B-B^prime.) Dashed lines (1) are time lines connecting evaporite minima beginning cycle 12 and ending cycle 13. Solid line (2) connects the first and last bed of anhydrite or halite in the two cycles.

End_Page 88-------------------------

3. The aggregate sulfate thickens outward from 5 to 57 ft (1.5 to 17.4 m), the rate of 1.3 ft/mi (25 cm/km).

4. The aggregate salt thins from 64 to 0.1 ft (19.5 to 0.03 m), the rate of 1.65 ft/mi (31 cm/km).

5. The two major cycles are recognized easily in cores, and are distinguishable on logs, in samples, and on drilling-time records from well to well across the 39 mi (63 km), in spite of change in facies.

Seemingly, space in the depositional basin was filled with synchronous facies appropriate to salinity gradients and map positions of those gradients. As the sulfate facies is only slightly thinner than the halite facies, foot-for-foot deposition of halite and sulfate rocks seems to have occurred "simultaneously" across a distance of 38 miles (61 km).

Cross section A-A^prime (Fig. 5) illustrates that the most soluble rock type in each cycle forms a lens within a larger lens of the next most soluble rock. Taken collectively, all

Fig. 13. Highly speculative correlations of two cores 130 mi (209 km) apart. Distance between maxima of cycles 12 and 22 is about 168 ft (51 m) at Midland; the stratigraphically equivalent(?) section at Sarnia, Ontario, is about 160 ft (49 m) thick.

End_Page 89-------------------------

the salt beds shown in cross section A-A^prime (Fig. 5) show the south half of a gross lensing pattern or wedge, as do the terminal edges of the sulfate beds, yet time lines are nearly parallel.

The outstanding fact to be learned from study of the Detroit River evaporites is that the evaporite maxima of many cycles can be correlated. This means that maximal concentration of brine in any cycle was reached at basically the same time in all places open to sea water in that cycle, and that lines constructed to pass through maxima and minima of the cycles are time lines. Evaporite rocks deposited within any cycle are of many types, which occur as sequential lithofacies forming wedges or lenses that thin out of the basin. The time lines connecting maxima and minima are very nearly parallel and these lines cross convergent lithologic boundaries as rock types change laterally.

With essentially-parallel time lines demonstrated, lateral change in rock types becomes evidence of synchronous deposition of similar thicknesses of halite, sulfate rock, and carbonate rock. Slight thinning of the section is attributed to gentle paleoslopes and almost-uniform subsidence.

Regional Influence of Cycles

As a depositional cycle moves toward the time of maximal salinity, the salinity gradients will have moved outward from the center of the basin. Increases in salinity mark the separate transgressions of brine from basin center. Of course, gradients regressed as water freshened during the dilution phase. The depositional edge of an area influenced by the effects of a series of evaporite cycles might show evaporitic maxima by thin beds of dolomitic limestone in limestone, or by fine-grained, non-fossiliferous limestone in fossiliferous rock. The influence of evaporitic conditions seems to have gone far beyond the basinal boundaries normally shown on the maps. Gardner (1974, p. 39-40) considered the Middle Devonian anhydrite beds of northern Indiana to be laterally equivalent to the Horne member: "The maximum lateral expansion of Lucas anhydrite deposition into Indiana corresponds to...maximum salt extent..." illustrating "the migration of peripheral anhydrite facies in unison with advance environments in which halite was depositing." The evidence at Midland (cross section A-A^prime) would suggest that the cycles likely to have had influence at a great distance from the basin depocenter would be cycles higher in the section than cycles 7 and 8, picked by Gardner. The most far-reaching cycles probably were the salt-bearing cycles 12, 13, 14, 15, 17, 20, and 21.

Correlation between the Dow 27P well at Midland, Michigan, and The Dow Chemical Company of Canada 43 Salt well near Sarnia, Ontario (Fig. 13) is based on the assumption that beds of halite and anhydrite at Midland are facies of anhydrite and dolomite at Sarnia. The Massive salt (cycles 12 and 13) comprises six cycles (at Midland there are at least four cycles) and the maxima of cycles 15, 21, and 22 are each represented by beds of anhydrite or gypsum only a few feet thick. If these speculative correlations prove to be correct, then time lines represented by evaporite maxima of cycles 12 through 22 can be shown to extend across the eastern half of the Michigan basin, bracketing lithofacies that change greatly, but the section thins only about 8 ft (2.4 m) in a total section of about 16 ft (51 m). This correlation across 130 mi (209 km) is conjectural, but near Midland, similar changes in cycles 12 and 13 have been demonstrated--with good control--across a distance of 38 mi (61 km).

Conclusions

Maxima and minima of many of the evaporite cycles in the Lucas Formation (Devonian) of Michigan can be correlated across moderately long distances. Depositional environments changed gradually in a cyclic manner that is evident both vertically and horizontally in the stratigraphic section.

The unbroken vertical succession of evaporite lithofacies is judged to be normal and interrupted cycles are considered abnormal. Maxima and minima of individual correlating cycles define time lines that are nearly parallel, and the evaporite rocks deposited between time lines connecting minima (from one set of rocks that indicate freshening of brine to the next) I regard as synchronous. Because lithofacies tend to

End_Page 90-------------------------

thin from the center of the basin in long, narrow wedges, and because time lines tend to be parallel, there are numerous instances of lithologic boundaries that cross time lines. Most cycles show reversed sequences with "inverse" or "recessive" evaporites in the upper half of the cycle, so that the lithologic boundaries tend to be inclined, at both top and base.

I believe that available depositional space was filled with synchronous facies appropriate to salinity gradients in the parent brine and map positions of the gradients with time. Halite and sulfate rocks apparently were deposited simultaneously and in almost-equal thicknesses across considerable distances. Evaporites that grade outward into less-soluble facies suggest deposition in one large body of water. Absence of potassium salts and high-bromine halite requires that the mechanisms of deposition moved potassium and other bittern ions out of the basin; one such means would have been by reflux bottom currents. Reflux currents require that the brine level was only slightly less than normal sea level outside the restricted basin, to bring about a continuous flow of ions from the basin; lack of bitterns precludes massive drawdown.

The apparent filling of available space suggests that each cycle may have involved alternations of water depth ranging from a few inches, i.e., a "filled" basin, to depths equaling thickness of rocks in individual cycles. During a few cycles, water may have been as deep as 60 ft (18 m). The gradual transition of facies laterally and vertically may have been the result of advances and retreats of a sabkha-playa environment. However, I suggest that the widespread evaporite strata of the Lucas Formation were more likely subaqueous, perhaps partly deposited in very shallow water, and mostly within a large body of water.