Nonetheless, Russell. {1968), Larson and Pitman (1972), Hays and Pitman (1973), Turcotte and. Burke (1978) and Harrison (1980) have used the age versus ...
Marine Geology, 58 (1984) 373--400
373
Elsevier Science Publishers B.V., Amsterdam -- Printed in The Netherlands
GLOBAL TECTONICS AND EUSTASY FOR THE PAST 2 BILLION YEARS
THOMAS
R. W O R S L E Y ' , D A M I A N
N A N C E I and J U D I T H B. M O O D Y
2
'Department of Geological Sciences, Ohio University,Athens, O H 45701 (U.S.A.) 2 0 N W I - Battelle, 505 King Avenue, Columbus, O H 43201 (U.S.A.) (Received July 5, 1983 ;revised and accepted October 18, 1983)
ABSTRACT Worsley, T.R., Nance, D. and Moody, J.B., 1984. Global tectonics and eustasy for the past 2 billionyears. Mar. Geol., 58: 373--400. Continental freeboard and eustasy, as gauged by the relative position of the world shelf break with respect to sea level,have varied by + 250 m from today's ice-free shelf break depth of ~ 200 m, during the past 600 Ma. Assuming constant or uniformly accreting continental crust and ocean water volume in an ice-free world, sea level fluctuations can be attributed to variation in the world ocean basin volume caused by changes in either its area or its depth relativeto the world shelf break. A n increase in volume and lowering of sea level occur as: (1) the world ocean floor ages, cools and subsides; (2) accreting continents collide, thicken and decrease in area; and (3) poorly conductive continental platforms become thermally elevated due to a size-induced stasisover the mantle. Conversely, a decrease in the age of the world ocean floor, attenuation of continental crust during rifting,and an increase in continent number and mobility, will reduce the world ocean basin volume and raisesea level. Theoretical sea level calculated from these principles correlates well with calibrated, first-order cycles of eustatic sea level change for the Phanerozoic. The record closely fits a simple model of retardation and acceleration of terrestrialheat loss during alternating periods of supercontinent accretion and fragmentation. Calibrated to sea-levelhighstands, successive first-ordermarine transgressionsand orogenic "Pangea" regressions characterize a self-sustaining,~ 440 M a plate tectonic cycle for the late Precambrlan and Phanerozoic. The cycle can be recognized as far back as 2 Ga from the tectonic evidence of continental collision and riftingrecorded in global orogenic peaks and mafic dike swarms, and may be related to major episodes of glaciation and evolutional biogenesis. INTRODUCTION T h e growing interest in long-term (i.e., ~ 1 0 s yr) episodicity in tectonic processes (Stille, 1924; Sloss, 1963; Holmes, 1951; Vinogradov and Tugarinov, 1962; Burwash, 1969; Condie, 1976, 1982; Fischer, 1981, in press; Anderson, 1981, 1982; Mackenzie and Pigott, 1981) has important consequences for terrestrial heat loss and the differential terrestrial heating that ultimately powers plate motion. A s heat production is a continuous process w h o s e rate is smoothly declining through time (Fig.l), demonstrated plate tectonic episodicity and its effects o n the hydrosphere and biosphere m u s t be manifesta0025-3227/84/$03.00
© 1984 Elsevier Science Publishers B.V.
374
4
A
1
t~
~--o~:
Model 2 3
2 Upper Mantle Lower Mantle Convection Begins
0!
I
4
I
I
3
2
I
1
I
Present
Ga
Fig.1. Heat flow (HFU) at the earth's surface through time (Ga). Model 1 is for an earlyconveeting earth (MeKenzie and Weiss, 1975) and Model 2 for a delayed-eonvecting earth (after Lambert, 1980). Note that both models are virtually identical for the past 2.5 Ga, showing a smooth exponential decline o f heat production.
tions of variable terrestrial heat loss. Furthermore, that this episodicity is at least in part a deterministic (nonrandom) process is demonstrated by the periodic recurrence of two and probably three Pangeas in the past one billion years (Windley, 1977; Condie, 1982). Given the insulating nature of continental tectosphere with respect to subcrustal heat flow (Anderson, 1981, 1982), clustered continents (Pangeas) should retard heat flow and therefore promote hemispherical asymmetry of terrestrial heat loss. Oscillatory terrestrial heat loss would therefore appear fundamental to the repeated assembly and fragmentation of supercontinents. Figures 2--5 illustrate recent attempts at compilation and synthesis of long-term quasi-periodic tectonic episodicity and form in large part the data set u p o n which we base our model of periodic Pangeas. The authors of these syntheses agree that the driving mechanism of tectonic episodicity and its geological, climatological, and paleontological consequences are liv_ked to episodic changes in heat flow patterns and ~ the cause of the changes to be variations in the intensity and style of convection patterns in the mantle. Anderson (1981, 1982) takes the reasoning one step further and offers evidence that the positions of the continents themselves are the cause, n o t the result, of changing mantle convection patterns. In this article we use
375
L
I
,I
Antarctica Shie,~ Indian
~
/
'astern Ghats
~
Shield
Australian I
P a ~ rican
~
I
~
Kibaran I Mo~biquian Kibaran I |
Eburnian ~ Lirnpopo I~
I
C~r~n
Trans-Amazonian
~
~
Stanovoi ~
Dalslandian I I Grenville I I 0.5
I
1.0
,~
,
Karelian ~
1.5
ShieId
Brazil-Guiana
Sh,eIds AIdan
Shield Ukrainian Sh,e,d
Baltic Shield Canadian-Greenland
Elsonian Hudsonian ~ ~ I
Shield
Central African Shield SouthAfrican
ShieIds
I
2.0 Age, Ga
Fig.2. Major orogenic periods 0.5--2.2 Ga, after Condie (1976). Note that peaks ending at 1.9 and 1.6 Ga are not clearly resolved.
the above information to demonstrate that relatively simple relationships govern the quasi-periodic behavior of the earth's tectonic history. To reconcile the repeated formation of Pangeas, we will first summarize the factors that control continental freeboard, or the relative elevation of the continents with respect to sea level (Wise, 1974), and then outline a model that accounts for hemispheric, quasi-periodic acceleration and retardation of heat loss. The model produces a heat~iriven, self~eplicating cycle of clustered and scattered continents with consequent first~rder effects on the freeboard, and hence climate and biogenesis (irreversible jumps in biotic complexity) consistent with available geologic data outlining plate motions for the past 2 Ga. Finally, we calibrate the model to the geologic record using the history of first-order eustatic sea-level changes. One of the most attractive features of the model is its ability to derive global geologic history on a ~ 50 Ma resolution using only plate tectonic and heat flow information. The averaging techniques used to construct the model recognize local or short-duration events only in proportion to their global contribution to the geologic record. Therefore, the model in its current version cannot be expected to represent, necessarily, either the geologic
376 100 8o
50
20 -Cumulative Percent of Present Day Levels
11 -10
Percent Present Atmospheric Level 0 2 i
2.0 Characteristic Crustal Features and Events
I
I~
1.0 Time Before Present (Ga)
J
Orogenies
i I
Biogenesis Proposed Tectonic Phases
I- - - I I III Intra-Plate Tectonics
* EUcsryote ,I + M~ll=oa 1.3 Ga I I =I~ IV _I Macro-Plate (Benioff Zone) Tectonics
Fig.3. Crustal controls and trends for the past 2 Ga (modified from Goodwin, 1981). history of any given crustal block or the detailed history of freeboard in general. However, residual differences between predicted and measured values of eustatic sea level probably represent real second.order geologic events (Fig,5), such as continental glaciation or continent, to-continent collision. TECTONIC CONTROLS OF SEA LEVEL Continental ~ b o a e d r ~ n t s the complement to eustas¥ and both are most readily ~ d as ~ w~ depths at the world ~ f ~ m k . The posit/on of the shelf break ~ t t e n t t y p r o ~ s a ¢onveniw~:re~e~qmce by which to gauge ~ o n i c c o n ~ l s of sea level. In ~ convention, a drowned
377
CLIMATE
Ice Sheets
I
600
Ice Sheets
I
Is,.I Oe. [
]
400
1',. Ju,.I Cret. I C'n. I
200
0
Ms Fig.4. Outlines of Phanerozoic history (after Fischer, 1981, in press). Sea level curve from Vail et al. (1977), granite emplacement curve from Engel and Engel (1964). world shelf would have a negative freeboard as the result of a positive eustatic sea level. However, for the purpose of avoiding confusion, we adopt the u n o r t h o d o x convention of quantifying negative freeboard such that values for both freeboard and eustasy are positive when shelves are flooded and negative when t h e y are emergent. Currently, the negative freeboard for the average world ocean is + 130 m (or + 200 m for an ice-free world if isostatic effects are n o t taken into account). However, negative freeboard has oscillated by more than 500 m during the last 600 Ma from deeper than +500 m (300 m above its present ice-free value) during the Late Cambrian/ Early Ordovician and Late Cretaceous, to depths less than zero (shelf break emergent) during the latest Precambrian and Early Permian (Fig.5). Freeboard changes as a function of the volume of water in the world ocean, and the volume o f the world ocean basin itself. The main mechanism for changing the volume of water in the ocean is to sequester a portion as ice caps on continents, a process that can lower sea level by as much as 200 m (Pitman, 1978). We will return to water-volume effects when we~ r disCuss second~)rder feedbacks to the primary factor of basin volume. The volume o f the world ocean basin is controlled b y three geometrically distinct b u t tectonicaUy interdependent mechanisms (Fig.6): (1) the mean elevation of the world's ocean floor with respect to the geoid; (2) the continent/ocean basin areal ratio (or simply the area of the continents); and
378
Changes of Sea Level Ridge Volume l Rising in Meters Falling 300 0i ' ,I 30(~ High Low II l I * I ...., ....
Periods 0-
•
Tertiary
100--
C retaceous
200--
Jurassic Triassic
fting
Permian 300-- Pennsylvanian Mississippian Devonian Silurian ,oo_
ii__
Ordovician
500-
Cambrian 600--
Precambrian
Present Sea Level
mg
t |
700--"
I
%~lb
I
II
Fig.5. Comparison of global sea-level changes and mid-oceanic ridge volume. Sea-level curve is after Vail et al. (1977), as calibrated by Pitman (1978)and adopted by Vail and Hardenbol (1979); the Vail et al. first'order curve, which averages out the sharp secondorder sea level drops shown in this figure, is plotted on Fig.14. Ridge volume is from Mackenzie and Pigott (1981) and Shanmugam and Moiola (1982). I I
'
~
I
I I
t I
' I
Ocean
i
,
~
',
I
b
i
Vo,ume
. . .
;'~':"::-':::::'~';';'~~'~'~
'~';;?~ ~ ? 1
~" " I I
,
b)
Mantle
~I I
I
I
I
~
,
::
-
Mantle
I I
,
Fil.6. Controls of ocean bidn volume. (a) Tbexmaily controlled blthymetzy of the ocean floor; (b) then'nally con~oOed continental elevation; (c) orogel~l~a]|y ~ t x o i l ~ continental compre~ion Continental compression will decrease continent~ area and hence i n crease ocean volume, Conversely, continental extension d u ~ ~ftlng will decrease ocean volume
379 (3) the mean elevation o f the continental blocks with respect to the geoid. Each of these controls on volume of the world ocean basin will be discussed in the following sections. Ocean f l o o r elevation
Sclater et al. (1971) have shown that oceanic lithosphere thickens, cools and subsides with increasing age. Davis and Lister (1974) have demonstrated that the depth increases linearly with square r o o t of age and Parsons and Sclater (1977) have fit the age versus depth relation to the equation: d ( t ) = 2500 + 350 t ~/2
(1)
where d = present-day b a t h y m e t r y in meters and t = age in Ma. As we are concerned with the vertical distance between the ocean crust and the shelf break, we subtract 130 m to yield: d ( t ) = 2 3 7 0 + 350 t ~/2
(2)
From this equation the seafloor will obviously subside if the world ocean crust ages and shoal if it becomes younger. F o r example, all other factors held constant, the shelf break to seafloor depth would be 4845 m [2370 + 350 (50) ~/~] for 50 Ma ocean crust (Fig. 7a) b u t would subside to 5081 m
,b)
d: s 81
Fig.7. Bathymetry of 50 Ma (a) and 60 Ms (b) ocean floor (Parsons and Sclater, 1977) and consequent increase in ocean volume assuming fixed continental blocks.
380
as the world ocean floor ages to 60 Ma [2370 + 350 (60)1t2], yielding a 236 m drop in sea level (Fig.7b). In practice, numerous assumptions are tied to these numbers and the effects o f loading and the extent to which oceanic and continental lithosphere are decoupled, are ignored. Nonetheless, Russell {1968), Larson and Pitman (1972), Hays and Pitman (1973), Turcotte and Burke (1978) and Harrison (1980) have used the age versus depth relationship to correlate high sea level with young world ocean floor. World ocean floor can become younger either by increasing seafloor spreading rates while retaining constant ridge length or by increasing ridge length while retaining constant spreading rates or a combination of both, as either serves to increase ridge volume and hence raise sea level. As Pangea breakup increases ridge length (see Southam and Hay, 1981, for a review), we start with the assumption (as do Berger and Winterer, 1974) that seafloor spreading rates remain largely constant. Assuming uniform spreading rates equivalent to today's, Berger and Winterer (1974) have calculated the age of the world ocean crust as a function of the breakup of a Pangea. During such an event, a 100% "Pacific-type" world ocean bordered by active margins would evolve into one with a uniformly increasing proportion of "Atlantic-type" ocean bordered by passive margins. We have adopted their general model here (shown schematically in Fig.8) but have calculated ocean-crust b a t h y m e t r y using our modified d ( t ) = 2370 + 350 t i n relationship instead of their empirical age-depth curve. We can see from Fig.9 that the Berger and Winterer world ocean has a mean age of 53 Ma and shelf break to seaftoor b a t h y m e t r y of 4918 m during Pangea, producing an ice-free shelf break depth of +340 m. However, Pangeas are known to have had minimal, if any, shelf break flooding (see Fischer, in press; Sandberg, 1983). Thus, we conclude that seafloor elevation alone cannot be the sole control over water depths at the shelf break and that either or both continental area and continental elevation
World Seafloor = 53 Ma
" A t l a n t i c " = Rifts
X W o r l d S e a f l o o r = 4 8 Ma " A t l a n t i c " 80 Ma Old
X W o r l d S e a f l e o r = 59 Ma " 'Atl anti c " 160 Ma Old
Fig.8. Schematic representation of the Berger and Wmte~er (1974) fragmenting Pangea model: (a) a Pangea ready to fragment surrounded by 53 Ma Pant~alassa; (b) the system 80 Ma after rifting. " A t l a n t i c . t y p e " oceans of 40 Ma average age represent 17% of the world's sea floor. The weighted average of " A t l a n t i c " and "Pacific,type" ocean floor yields a world's sea floor of 50 Ma; (c) the system 160 Ma after r i f t i n g ; " A t l a n t i c - t y p e " oceans now average 80 Ms and represent 33% of world ocean floor yielding a world seafloor age of 62 Ma; this stage illustrates today,s world ocean. (In reality, the South Atlantic and "passive Indian" oceans opened later than the North Atlantic so that today's 33% "Atlantic-type" oceans have an X of 70, not 80 Ms, yielding a world seafloor age of 59 Ma. We used the 70 Ma age to calibrate Fig. 9.)
381
Mean Host Flow (HFU) 0 81$
Time ElapsedAfter Pongee Breakup(Me) 80 160 240 I I 1
320
.=
1.8-
7 0.6-
2.1-
.~ u
;=6(]-
-Today's World~t~ . . . . . .
-
.
~
!
O! 55 . . . .
" 2.0-~=
. 0,6-cl
O 2,4-
~°
2.3-
~ 00.6--
o 2.8--
~ ~; 4e 0
50.8!
_ / : -
i r- -Pangea.
I --
l 11
__/ .
.
.
--.200 ~
o
-f-At~'ieotic"Z--~r~/2 Closed 4989 0
. .
.
"-
MaximumFlooding I I ('ATL;' 1/2 Open) 22 33 22 11 % "Atlantic-type"/OceanFloor /
OpeningPhase-~-~--Closing
"-
4918 ~
~--+2690 ~ +34o
47g§_~ I--+463 | .~ 0
