A reply to E. C. perry\'s comments

Geochrmm et Cmmchrmica Copyri&t 0 1990 Pergamon

0016.7037/90/$3.00

Am Vol. 54, pp. 1181-l 184 Press pk. Printed in U.S.A.

+ .oO

REPLY

A reply to E. C. Perry’s comments S. EPSTEIN’ and J. KARHU*

’ Division of Geological and Planetary Sciences, Califi~inia Institute of Technology, Pasadena, CA 91 125, USA * Geological Survey of Finlarf 9,SF-02 150 Espoo 15, Finland (Received October 3 1, 1989; accepted in revisedjimn January 30, 1990)

As A BASIS FOR OUR RESPONSE to PERRY ( 1990), we wish to present a short resume of the basic premise of the KARHU and EPSTEIN (1986) paper (referred to henceforth as KE 86). This resume will assist the reader in evaluating our replies to Perry’s comments. The 6180 values for two coexisting natural minerals, such as quartz and apatite in isotopic equilibrium, when plotted on a linear graph, can provide information regarding the conditions under which the quartz and apatite were formed. Figure 1 shows this plot. Point X represents the 6’*0 values of two minerals, p and q (i.e., quartz and phosphate), formed in an ocean whose 6”O is O%Ounder isotopic equilibrium at a temperature T. At this temperature, AqmD= 6’80,-6’80, = 15%~ This difference at T is independent of the 6180 of the water or of the minerals, as long as they are in isotopic equilibrium. The line T3 is an isotherm of T. The A4_pvalues at higher temperatures will be lower than 15 and approach O%Oat very high temperatures. The 45” line designates the T = m isotherm. Other intermediate temperature isotherms are represented by the parallel lines between T3 and T, . If mineral pairs p and q form at various temperatures warmer than T in the ocean where amount H20 @ amount q + p and the aI80 of the ocean water is O%a, the 6’*0 plot will form a line labeled Z. The slope of this line is a function of the change of the ratio 1000 In apH20 to 1000 In c+Hzo with temperature, where 1000 In (Yis approximately the same as ~‘Qn,“,,, - 6’80H,o. All the above statements are rigorous and are governed by the fundamental oxygen isotope fractionation factors for the system q - p - H20. Generally, once the relationships between cy and temperature are calibrated, it is possible to calculate the temperatures at which the minerals formed and the 6’*0 of the interacting water. Isotope data for q - p pairs which were not formed under the specified conditions will give points which do not fall on line Z. The only cautionary note necessary is the possibility that a non-equilibrium pair may fortuitously acquire 6’*0 values which may result in a data point on line 2. However, it is highly unlikely that a series of 19 points obtained for microcrystalline quartz (cherts) and coexisting phosphates, formed at different ages and collected from different parts of the world, will coincidentally fall on the rigorously defined line Z, unless they are equilibrium pairs in waters whose 6”O are in the neighborhood of the isotopic composition of the same media. If a q - p pair is not in isotopic equilibrium, then the data point can fall anywhere, including outside the area bounded by the isotherms T, and T3. Isotopic re-equilibration with

excess water whose 6”O is lower than O%Owill result in 6’*0 points below the Z line. Dry re-equilibration between q and p will result in data points to the left of or above Z, depending upon to what degree p or q is the dominant component. Figure 2 is a plot of A4_,, vs. 6”0,. On this plot, it is possible to observe the AQep values on a more sensitive scale. The isotherms are parallel to the x axis, i.e., constant Au_,, values. The slope of line V reflects the change of the ratio of 1000 ln aq_p to 1000 1n aq_&o with temperatures for a system Hz0 @ q + p and is equivalent to line Z in Fig. 1. The requirements necessary for the data points to lie on line V are the same as those that form line Z in Fig. 1. A straight-line relationship not defined by the above criteria (KE 86) will not correspond with the Z or V lines, and will bear no relationship to criteria for equilibrium or to marine temperatures for a system H20 & q + p. In Figs. 1 and 2, Phanerozoic-A samples illustrate the behavior of a suite of samples in isotopic equilibrium in a medium whose 6”O is approximately O%O,and PhanerozoicB samples are non-equilibrium q - p pairs and give data points randomly distributed on the graph. By independent criteria (i.e., granular microcrystalline quartz), PhanerozoicA samples should be the “primary” chert, and the enclosed biogenic phosphate should be isotopically equilibrated with the water medium. In his Comment, PERRY (1990) does not accept: (a) that such rigorous criteria can be obtained from the oxygen isotope systematics and (b) that well-preserved cherts which are microcrystalline quartz retain their original isotopic composition. According to Perry, these samples have mostly been drastically modified by late diagenesis with fresh water and by high-temperature metamorphism. Our response to specific points raised by Perry. in the designated paragraphs of his Comment, is given below: Paragraphs 2 and 8. Perry refers to the paper by VENGOSH et al. (1987) which shows that some of the samples of the Mishash formation in Israel which contained both phosphates and cherts have undergone various degrees of late diagenesis, with low 6”O waters. The original records of these cherts were modified but the phosphates retained their original marine values. Thus the Aumpvalues for these samples ranged from 18 to about 2%0 and plot against 6180, linearly with a slope of 1. As specifically pointed out by VENGOSH et al. (1987), the linear array arises artificially because 6’*0,, is essentially constant while the 6’*0, varies in both variables. This is not the case for the Phanerozoic-A samples that KE 86 analyzed. The 6’*0, 1181

values vary sympathetical1.v

with the ~5~~0, values

118’

5. l:psteln and J Karhu

phutes (SZO-+JOper mil). The phases ofthis assnnblagc

car be interpreted as all having been deposited with a late fresh water fluid.” This is the criteria that KE 86 used fin isotopic non-equilibrium between the cherts and phosphates. The> also state (p. 246): “It seems significant that thL>fresh-watrr assemblage does not include an apatite member: 6’“O in phosphates seem to be sufficiently immune to the diagenetit recrystallization process.” This demonstrates the resistance of phosphate to isotope reequilibration. In the Mishash Campanian Formation. there is no doubt about rhe ability to identify the well-preserved and the late diagenetically altered material. The authors have established three important cnteria which coincide with those used by KE 86

FIG. 1. A plot of the 6’*0 values for coexisting quartz and phosphate in isotopic equilibrium. The lines parallel to T3 are isotherms at various temperatures, and line z represents the 6’80,-S’80, plot for a system HrO-q-p where the amount Hz0 % amount of q + y. The data points are the data from KARHU and EPSTEIN(1986). The filled in circles are for equilibrium 4 - io pairs and open circles are for nonequilibrium pairs.

(see Fig. 1 and Fig. 5, KE 86), a major point in their paper. Variations of the Ay_* vs. d”O, (Fig. 2) cannot therefore be interpreted in the manner suggested by Perry. This array is readily interpreted as a temperature variation. not as a consequence of the effects described by VENCOSH et al. (1987). For these reasons, Perry’s discussion in his paragraph 8, and subsequent plotting of the VENGOSH et al. (1987) line over the KE 86 data in his Fig. 2, is completely inappropriate and misleading. In addition, Perry’s regression line for the KE 86 data in his Fig. 2 is badly distorted by the inclusion of the data for the light-water sample in his calculation. A comparison of Perry’s regression line with the Fig. 2 of this Reply demonstrates this discrepancy. In actual fact, the paper by VENCOSH et al. (1987) lends excellent support: (I) for the choice of criteria by KE 86 for the preservation of the isotopic records in the chert-phosphate pairs, (2) for the validity of the isotopic temperatures they record, and (3) for the general approach taken by KE 86. VENGOSH et al. (1987) studied a Cretaceous (Campanian) carbonate-chert-phosphate marine deposit whose moderate temperatures were well defined from their well-preserved carbonate marine fossils. Their conclusions are summarized in their abstract. Item (b) in their abstract states “an early diagenetic assemblage formed in equilibrium with a marine solution is recorded in calcite of micrite cement (+26 to t-27.5), silica in opal-CT (31.5 to 33.5), silica in quartz in homogeneous chert (i-29 to +32), silica in quartz asfiugments 111chert hreccius (+31 to i-33 per mil) and some of the quartzose matrix in silicified phosphates. The italicized statements are relevant to the KE 86 Phanerozoic-A samples which are homogeneous microcrystalline quartz. In addition, these authors state (p. 241): “In a quite general way, our results confirm the above notions: the isotopic imprint of all homogeneous cherts is practically indistinguishable from a seawater signal.” This is what the KE 86 paper infers. Item (c) states “a later diagenetic assemblage contains calcite spar, intilling of fossils (+20 to +26.6?&), silica in matrix of chert breccias and coarse quartz matrix in silic$ed phos-

Granular microcrystalline quartz (cherts) arc early diagenetic deposits reflecting the temperature and isotopic composition of the interacting fluids which. m effect, may be ocean water. Coarse and other recrystallized forms of quartz have 6’“O records which are fixed by post-early depositional processes, including metamorphism. The phosphates are more difficult to exchange and, thus, the conodonts may be good candidates for recording reliable oxygen isotope paleotemperatures.

Purugraphs 3-5. Perry uses the three Triassic samples repeatedly to argue that KE 86 used samples that were nonmarine, metamorphosed, and misdated. The petrified wood sample (P41) is possibly non-marine. but the coexisting phosphate-quartz data suggest silicification in a water with 6”O near zero. This result suggests that the depositional environment of the Chinle Formation at the collection locality was marine, coastal, or near-marine. The temperature is like the other Triassic samples. We see no reason to throw out this result. Perry dismisses the result for sample P58. in part because of the possibility that a supratidal horizon may overlay the sample horizon and evaporite fluids may have been the silicifying fluid. If such speculative, undocumented environmen15

I

I

I

_~____~__

1

n

PHANERXOIC,LIGHTWATEP

l

PHANEROZOlC,A

0 PHANEROZOIC,B

ISOTHfRM

.

/'

/ ,.' c

/ / //

FG. 2. A plot of the ~3~~0,- S’“0, (&,) vs. 6’80,. Line V is similar to line z of Fig. 1 in that it represents the q - p plot far a system where the amount of Hz0 % the amount p t q. The data are as described in Fig. 1.

Reply tal interpretations are used to disqualify samples, then no samples are suitable for geochemical analysis. The 6”O of this chert is among the lowest of the Phanerozoic samples, suggesting formation in the presence of diagenetic meteoric waters, not 6”O enriched evaporite fluids. This is totally consistent with the observation that 6”O of the chert is lower than that of the host carbonate. Perry’s suggestion that metamorphism preferentially lowered 6”O of the chert relative to the carbonate is highly improbable; carbonate is much more susceptible to isotopic alteration than quartz. Additionally, this unique metamorphic alteration would have to have occurred without coarsening of the chert grain size. The KE 86 interpretation is clearly a simpler one. Perry rejects results from Triassic sample P49 on the grounds that the age has been reinterpreted as Permian by SPEED ( 1977). This sample comes from a structurally complex region containing a variety of volcaniclastic sediments and cherts of problematical origin. SPEED (1977) has indicated that “part” of this complex is probably Permian, but there is no specific age information on the extraordinary chert beds west of Mina, Nevada. Perry’s assertions about the age of P49 are overstated and misleading. The isotopic data for P49 in KE 76 (KNAUTH and EPSTEIN, 1976) and KE 86 are consistent with other Triassic samples, suggesting that the age of the chert is Triassic rather than Permian. The interesting possibility, however, that chert ages can be argued from stable isotope measurements must await further analyses of many more samples. The distinctions between Permian and Triassic data are still sketchy. According to Perry, evidence from nearby conodont “likely” indicates some metamorphism of KE 86 samples. Actually. in our conversations with Dr. Harris, the state of the conodonts varies with the local conditions and cannot be used to characterize the temperature conditions of a general area. The 6’80 of the conodonts which KE 86 analyzed (Table 2, KE 86) show no isotopic evidence of high-temperature metamorphism. They give calculated temperatures, assuming equilibrium with marine waters of 6180 = - 1%Othat range between 30 and 70°C for post-Jurassic samples, including two Triassic samples from Nevada. However, if we allow a value of -2%0 for the oceans in the post-Jurassic, then the conodont calculated temperatures are more in line with those calculated from A4_p data. As will be discussed later, a -2%0 value for the aI80 of ocean water is not entirely out of the question. In addition, if the conodonts were metamorphosed with low 6180 water and/or exchanged at 200-3OO”C, the 6l”O of the conodonts would be lower than those that KE 86 report. For example, some of the Precambrian phosphates have values of 8 and 7%0 which give temperatures ofjust over 100°C. Perry’s contention that the chert-phosphate pairs of some of the Phanerozoic-A samples have been metamorphosed has little merit. Pclrugruah 6. Using equations of KNAUTH ( 1973) and KE 76, Perry calculates the 6”O of the “original unmetamorphosed” cherts which were formed in ocean water whose 6”O is O%O.For example, in the case of sample P58, the 9%0 correction simply states that the 6”O of the water in which P58 was formed was approximately -9%0. In the case of the other Phanerozoic-A samples in Fig. 2, these corrections range between 1.2 and 5.1. These calculated 6”O corrections are

1183

not very precise. Errors of 2-4%o in the calculated 6”O values would not be surprising, considering the scatter of points which define the equations estimated by KNAUTH (1973) and KE 76. Perry then plots the corrections of the cherts against the 6”O of the phosphates in his Fig. 1, assuming that the 6180 of the phosphates were not modified by the post-depositional processes. It is not clear what Fig. 1 is supposed to show except that for the samples in question neither the quartz nor the phosphate formed in water whose 6”O is O%O.In actual fact, the chert and phosphates acquired their isotopic composition in equilibrium with water whose 6”O are calculated by KE 86. KE 86 have calculated the 6”O ofthe water in equilibrium with the twenty Phanerozoic-A samples, using a different approach. They utilized the relationships of 1000 In 01 with temperature for the q - HZ0 and q - p systems. The 1000 In LYfor these two systems are equal to 6”0, - 6’80H,o and 6”0, - S’“O,, respectively. For a given temperature calculated from 6”0, - 6”0,, one obtains the value of 6”0, - 6’80H,o for that temperature. The values for 6”0, and 6”0,, are measured and thus the 6’80H20 can be calculated. This was done for all of the twenty Phanerozoic pairs. Fourteen of the twenty samples gave 6’80H20 values between zero and -2.4%0. This group included the three Eocene, one Cretaceous, two Triassic, one Permian, two Pennsylvanian, all of the Mississippian, two Ordovician, and one Cambrian samples. A 6”0,,, value of -5.2%0 was obtained for one of the Cretaceous samples, -9.9%0 for one of the Triassic samples, and -5.2%0 for one of the Upper Ordovician samples. The other three samples gave 6180 values of -4.1, -3.1, and -3.4%0. There is no systematic trend with time observed for these aI80 values of the water. For an equilibrium system, the 6’80H20 values could have no bearing on the calculated temperatures for all these quartz-phosphate pairs since the temperatures could be calculated from the Ay-p values. For example, the calculated temperature for the Cretaceous sample which had a 6’80H,o value of -5.2%0 was I5”C, in agreement with the results of DOUGLAS and WOODRUFF (198 1) and VENGOSH et al. (1987). One of the Upper Ordovician samples had a 6’80,,o water correction of -5.2%0 and again gave a low temperature of 23°C. The Triassic samples had a 6”0,,, value of -9.9%0 which gave a temperature of -56°C. Actually, the calculated 6”O values of KE 86 for the samples P41, P67, P39, P49, and P58 are, respectively. -0.5, -3.1, - 1.5. -2.4, and -9.9. as compared to the respective values calculated by Perry of -1.2, -3.2, -3.2, -5.1, and -9.0. Both sets of values actually check reasonably well, considering the inaccuracy of the method used by Perry. There is only one way a set of samples, such as those comprising the Phanerozoic-A group, could give reasonable and consistent isotopic temperatures that are insensitive to the isotopic composition of the water in which they formed. The early diagenesis responsible for the formation of the quartzphosphate pairs took place under conditions where both the formation of the cherts, as well as the precipitation of the apatite by living organisms to facilitate isotopic equilibration, occurred concurrently. These processes could take place in water of dominantly marine origin, but which could also contain a meteoric water component. According to KNAWH

such early diagenetic fluids are usually coastal which represent mixed marine-meteoric ground waters which have 6’*0 values within a few per-mil of sea water. Such a model iscompatible with the KE X6 data and with the type ofcherts analyzed by KE 86, and probably give temperatures close to surfaces of the coastal oceans, rather than the deep-sea temperatures. Although these possibilities were implicit in the KE 86 paper, their use of the term “ocean water” may have caused some concern to Perry. Eventually, analyses ofchertsphosphate pairs will help in the dehneation ofdifIerences in the environments of chert formation and will contrihutc to a more detailed unde~tanding of the l~ecf~nislns of chert formation. In Iight of the above considerations, KE X6 concluded that the chert-phosphate pairs did not suffer metamorphism which altered the 6’% ofthe chert or ofthe phosphate, or introduce a foreign water component of different isotopic composition subsequent to the formation of the cherts. Perry’s presumption that the lowering of the 6’*0 of the cherts was due to the metamorphism which would destroy the temperature records is not valid. On the other hand, a good example of the consequences of metamorphism, or post-early alteration is well demonstrated in the Phanerozoic-B samples which Perry ignores. ~~~ff~ru~~ 7. KE 86 specifically state that the Triassic calculated temperatures are unusually high. They calculated temperatures for all of the Phanerozoic-A samples and have included in their papers sufficient precautions about their temperatures. &~ugrcrph 8. Perry’s argument about the possibility of a microcrystalline quartz being exchangeable hangs on the undocumented “many observers.” The one reference that he cites on this issue is the 1962 paper ofthe late professor Degens and Epstein. Actually, these authors did not present evidence for isotopic exchange in microcrystalline quartz. They cxpressed an opinion that this may be the most obvious explanation. In view of the knowledge at that time regarding the formation of micr~~sta]Iine cherts, this was not an unreasonable explanation, but further work on this subject was obviously needed resulting in papers by KE 76 and others. Perry’s other comments in his paragraph 8 have been dealt with in conjunction with paragraph 2. /Gragr~ph 9. The location of the “steep line” in Perry’s Fig. 2, as presented by VENGOSH et al. (I 987) in their Fig. 8, delineated the location of the late diagenetic samples in the Cretaceous Mishash Formation. Obviously. this line will be different for different geological times, geographical locations, and marine temperatures. For warm temperatures, this line may be located at the b”O, value of 22% intersection. Perry’s representation in Fig. 2 has no bearing on the PhanerozoicA samples and is generally misleading on how the “steep line“ is used. ~uqqz~h Il. Perry makes a blanket statement that all the Precambrian chert-phosphate pairs are metamorphosed and reequilibrated without presenting documentation. The ( 1979),

3.5 Ba Onverwacht samples contradict that conclusion. ‘1he 8”O of the four chert samples show a small range of 15.7 lo 17.5!%e, and yet the 6”O of the coexisting phosphate var! over a much larger range of&O to 17.9%. Thcst: are the rnoft metamorphosed of all the samples analyzed and should show the effect Perry invokes. They obviously do not show this effect. If reequilibration between coexisting quartz and phosphate were so readily attainable. why would such large inconsistencies between the fi”O exist unless au ohvrous lack of equilibrium exists in 3.5 Ba old samples’! ~~~u~~~lp~~12. We hope that the readers of the KL Xii paper will not ignore the warning hy the authors on the limitations of this preliminary e&t and wilt not use the temperature data in a reckless manner. Additional work. wilt undoubtedly improve and add upon the methods and establish additional criteria for the preservation of chert-phosphate pairs and, in this way, eventually arrive at precise. correct. and numerous temperature data which till tell us much about the Phanerozoic and Precambrian. We wish to add that we had the beneht of Pro&or Knauth’s wise counsel and major contributions to this reply, but we are solely responsible for this rebuttal This is Cont~bution No. 4847, Division ofGeo~o~ca1 and Planetary Sciences, California Institute of Techn~~logy. Pasadena. CA 9 I 125. USA. ffditorial

hand&:

Ci.

Faure REFERENCES

DEGENSE. ‘I’.and EPSTEINS. (I 962) Relationship between O”/CYh ratios in coexisting carbonates, cherts, and diatomites. Bull. .4mer. .4ssoc. Petrol. Geol. 46, 534-542. Douatis R. G. and WOODRUFFF. ( I98 1) Deep sea henthic foraminifera. In The Deep Sea (ed. C. blILIAN1).Vol. 7. Wiley Interscience. KARHU J. and EPSTEINS. (1986) The implication of the oxygen isotope records in coexisting cherts and phosphates. C&r&rim. ~~~srn~~i~n~ Ada SO, 1745-l 756. KNAUTHL. P. (1973) Oxygen and hydrogen isotope ratios in cherts and related rocks, Ph.D. thesis, California institute ofTechnology. KNAUTHL. P. (1979) A model for the origin of chert in limestone. Ci~o/ogy7, 274-271. KNAUTHL. P. and EPSTEINS. (1976) Hydrogen and oxygen isotope ratios in modular and bedded cherts: Geochim. Cnsmochim. Icta 40, 1095-I 108. PERRY E. C. (1990) Comment on “The implication of the oxygen isotope records in coexisting cherts and phosphates. Geochim. Cosmcjehim. Acta 54, I 175-l t 79 (this issue). SPEED R. C. (1977) Excelsior Formation, West Central Nevada: Stratigraphic appraisal, new divisions, and paleogeographic interpretations: In Paleozoic Paieo~e~~ru~~ qf the Western ignited States: Pac$ic C.‘oast Pa~eo~el~~~a~h.~ ~_v~~~~~~~F~ I. (ed. J. H. STEWARD).pp. 325-336. Sot. Ecnn. Paleo. Mineral., Pacilic Section, Los Angeles, CA. VENGOSHA., KOL~DNVY., and TEPPERBERG M. (1987) Multi-phase oxygen isotopic analysis as a tracer of diagenesis: The example of the Mishash Formation, Cretaceous of Israel. Chem. Geo/o~y 65, 235-253.