Geochimica et Cosmochimica Acfa Vol. 56, pp. 3281-3295 0016-7037/92/S%oO + 03 Copyright 0 1992 Pcrgamon Pres Ltd. F-rimed in U.S.A. Sedimentary carbonates through Phanerozoic time FRED T. MACKENZIE’and JOHN W. MORSE’ ‘Department of Oceanography, School of Ocean and Earth Science and Technology, University of Hawaii, Honolulu, HI 96822, USA ‘Department of Oceanography, Texas A&M University, College Station, TX 77843, USA (Received March 19, 199 1; accepted in revisedfirm January 15, 1992) Abstract-Plate tectonic processes play a critical role in the origin and distribution of sedimentary car- bonates through Phanerozoic time. The Phanerozoic age distribution of sedimentary properties like calcite/ dolomite ratio, inferred o&d and cement mineralogy, and survival rate of continental carbonates is cyclic. The cycles appear to be coupled to plate tectonic processes that give rise to global sea level change and changes in the properties of the ocean-atmosphere system. First-order changes in sea level are driven by the accretion of mid-ocean ridges: high accretion rate, high sea level; low accretion rate, low sea level. Although correlations between sea level and sedimentary carbonate properties are not strong, high sea level over an extended period of time appears to be correlated with low calcite/dolomite ratios, lack of inferred aragonite oiiids and cements, and maxima in the survival rate of continental carbonates. The opposite is true for extended periods of low mid-ocean ridge accretion rates and global sea levels. The lack of strong correlations may reflect an insufficient data base and the possibility of lags between sea level change and change in carbonate properties. Furthermore, the survival rate of continental carbonates appears to be affected by differential cycling and, therefore, may not be directly related to accumulation rate. It appears that the environmental conditions for early dolomitization and calcite oSid and cement formation are best met during extended times of high sea level when atmospheric CO2 levels are high and the saturation state of seawater with respect to carbonate minerals relatively low. During low sea levels, early dolomitization is less favored, and aragonite precipitates are more abundant because of low atmospheric CO2 levels and enhanced seawater carbonate saturation states. Differential cycling has modified the Phanerozoic sedimentary carbonate mass-age distribution. Because of erosion of younger units within continental carbonate cycles, it may be difficult to derive an unequivocal record of the partitioning of carbonate between the deep-sea and shallow-water realms of deposition during the Phanerozoic. This difficulty must be considered in further quantification of geochemical models describing the geologic history of atmospheric CO* and climate change. INTRODUCTION by various authors (e.g., GREGOR, 1970, 1985; MACKENZIE, 1975; GARRELSet al., 1976; VEIZERand JANSEN,1979, 1985; IN A PAPERpublished in 1969 ( GARRELS and MACKENZIE, VEIZER, 1988; GREGOR et al., 1988; WOLD and HAY, 1990). 1969), Bob Garrels and Fred Mackenzie initially presented Further compilations and interpretations of the Phanerozoic hypotheses for the meaning of the temporal variation in pro- sedimentary carbonate rock mass-age distribution have ap- portions of sedimentary rock types remaining today in the peared in the literature in the 1980s (RONOV, 1980; HAY, geologic column. In the book Evolution of Sedimentary Rocks 1985; WILKINSONand WALKER, 1989; WILKINSONand AL- ( 197 la), they further developed models of the sedimentary GEO, 1989; Boss and WILKINSON, 1991). In this paper, as a rock mass-age distribution and expanded on two hypotheses tribute to and in remembrance of Bob Garrels, we explore related to that distribution: ( 1) geochemical uniformitari- in more detail the concepts of geochemical uniformitarianism anism and (2) differential cycling. Geochemical uniformi- and differential cycling relevant to interpretations of the sed- tarianism implies that the total mass of sediments of all ages imentary carbonate rock mass-age distribution through Pha- existing at any given time in the geologic past may have had nerozoic geologic time. This study reflects one of Bob’s prin- about the same ratios of rock types that we observe today. A cipal interests in the later stages of his career. corollary to this hypothesis is that the fluxes of chemical con- stituents to the oceans have not varied greatly, at least during THE DATA BASE Phanerozoic time. Differential cycling implies that because of differences in relative erosional resistances or tectonic set- The calculations and interpretations of this paper are based ting, various components of the sedimentary rock mass cycle on data from a number of sources. There is still some dis- at different rates. This factor, along with diagenesis, may lead agreement in the literature concerning the best estimates of to differences in the ratios of rock types as a function of mass-age relationships within the major global carbonate res- geologic age in the sedimentary rock mass existing today. ervoirs (e.g., HAY, 1985; WILKINSONand WALKER, 1989). These concepts and others involving the sedimentary rock We will clearly indicate in the following discussions the data mass-age distribution have been explored and expanded on sources and our manipulations of the data base. To provide 3281 3282 F. T. Mackenzie and J. W. Morse Table 1. Phanerozoiccarbonate mass distribution. Muss Survival Rate Period Duration TOM TOM Total Calcite/ Total I-%S Total Logs TOtZ4 Logs (lO~-s) carbonate Dolomite Cslcite Dolomite Dolomite Dolomite C&i& Calcite CarboMtc CariYonatc (1~) (lO?oM) (lO%ms) Ratio flti y-l) (tons y’) (W~OIISy-l) (tons yz) (Wtons y-I) (tons y’) re&Iy 65 19.42 2.92 16.50 5.65 4.42 a.65 25.0 9.40 29.42 9.47 cretaceous 66 lo.48 3.88 6.60 1.70 5.88 8.71 10.0 9.00 15.88 9.20 consistency with the works of Wilkinson and colleagues timate because of the use of Hay’s CO* data (WILKINSON (WILKINSON and WALKER, 1989; WILKINSONand ALGEO, and WALKER, 1989). 1989), carbonate mass-age data will be given in units of 10 r3 For several decades it has been assumed that the Mg/Ca g Ca y-l. This convention introduces an overestimation of ratio of carbonate rocks increases with increasing Phanerozoic the cation mass of about 2% for each 10% of dolomite found rock age. An early portrayal of this trend in carbonate rocks in a rock mass interval (WILKINSONand WALKER, 1989). from the Russian Platform and North America is shown in The distribution of Phanerozoic System total sedimentary Fig. 2. The trend represents a general, but erratic, decline in masses with geologic age was obtained from the estimates of the calcium content and increase in the magn~ium content GREGOR ( 1985). The mass of carbonate rock in each system of these rocks with incising age (see VIN~CRAD~V and was calculated from the estimates of CO, found in Phanero- RONOV, 1956a,b; CHILINGAR, 1956). The magnesium con- zoic carbonates as given by RONOV ( 1980) and amended by tent is relatively constant in these carbonates for about 100 HAY ( 1985 ) . Dolomite and limestone abundances were cal- million years, then increases gradually. The magnesium con- culated using the GIVEN and WILKINSON ( 1987) compilations tent of North American and Russian Platform continental ofdata on the composition (Mg/Ca or MgC03/CaC03 ratios) carbonate rocks appears to increase at a geologic age that is of Phanerozoic carbonate rock samples. Table 1 gives the very close to, if not the same as, the age of the beginning of resuhs of our calculations. A tentative mass-age dist~bution the general increase in the Mg content of pelagic limestones of sedimentary carbonates and sandstones plus shales is given f 100 million years before present; RENARD, 1986). The do- in Fig. 1. The total carbonate mass makes up about 30% of lomite content of deepsea sediments also increases erratically Phanerozoic sediments in Fig. 1, perhaps a slightly high es- with increasing age back to about 125 million years before P 180% DotwaIte EOSO* . [ 0.40 u c j 0.30 a I 10 0.20 Z _______--___-_---_-AVHI~~w&on& rack ----------- RunrmPI&&ml I! O*‘O 0 8 !5 4 3 2 1 0 100 400 1000 zaoo Time (10'~) Age (104~BP) FIG. 1. Mass-age distribution of carbonate rocks and other sedi- FIG. 2. Magnesium to calcium weight ratios in Russian Platform mentary rocks plotted as survival rate (S) versus age. Total rock mass and North American carbonate rocks as a function of age. (Data data from GREGOR ( 1985) and estimates of carbonate rock mass from VINOGRADOVand RONOV, 1956a.b; and CHILINGAR,1956; from Table 1. figure modified from GARRELSand MACKENZIE,1971a). Geologic cycles of carbonate rocks 3283 B Sperber et al. (1994) MglCa=0.23 So C Schmoker el al. (1997) D Chlllngar (1996) MglCa=0.14 MglCe=O.lS Vinogradov and Ronov (1956b) F Given and Wilkinson(1987) Mg/Ca=0.14 MglCa=O.23 N=607 N=l7,353 Age (106 y BP) FIG. 3. Estimates of percent dolomite in Phanerozoic cratonic carbonate rocks as a function of age. (After WILKINK~N and ALGEO,1989). present ( LUMSDEN, 1985 ) . Thus, the increase in magnesium Table 2 and shown in Fig. 5. WILKINSON and WALKER( 1989) content of carbonate rocks with increasing age into at least developed mass-age models for these various carbonate res- the Early Cretaceous appears to be a global phenomenon, ervoir distributions similar to those used by GARRELS and and to a first approximation, is not lithofacies related. MACKENZIE( 197 la). In a later section, these mass-age re- Recently, the accepted truism that dolomite abundance lationships are discussed in detail. increases relative to limestone with increasing Phanerozoic age has been challenged by GIVEN and WILKINSON( 1987).