[A\lfRICAN JOURNAL OF SCIENCf, VOL. 290, NOVEMBfR, 1990, P.

1069-1089J

ESTIMATING ANCIENT SEDIMENT FLUXES CHRISTOPHER N. WOLD and WILLIAM W. HAY Department of Geology, Cooperative Institute for Research in Environmental Sciences, and Museum, Campus Box 250, University of

Colorado, Boulder, Colorado 80309

ABSTRACT. Original Phanerozoic sediment flux rates can be approximated by fitting an exponential decay curve to the observed mass/age distribution of sediments, then multiplying the ratio of the observed mass to the exponential decay curve mass by the zero-age flux rate predicted by the exponential decay curve. Applied to the mass/age distribution of existing Phanerozoic sediments, the reconstruction method indicates that original global sediment flux rates varied cyclically by a factor of up to three. This implies significant changes in the rates of uplift and mountain building and global plate tectonic activity.

INTRODUCTION

From an analysis of areas of outcrop of Phanerozoic sediments of North and South America, Gilluly (1969) recognized that the apparent decrease of the amount of sediment of greater age was a function of loss of older sedimentary material to erosion. As better data became available (Ronov and Yaroshevskiy, 1969), it became apparent that the volume/age distribution of Phanerozoic sediment has the form of an exponential decay (Gregor, 1970: Garrels and Mackenzie, 1971 a, b: Li, 1972; Veizer and Jansen, 1979, 1985: Veizer, 1988). Blatt and Jones (1975) estimated the area/age distribution on a global scale, using data on age and types of rock at 783 randomly determined latitude-longitude points on land. They concluded that the area/age distribution is well described by a decay curve with a half life of 130 my. The exponential decay relationship is so evident in data sets for areal, volurney', and mass/age distributions that Veizer and Jansen (1979) concluded that "the described exponential relationship is a fundamental law of the present day age distribution of geologic entities" (p. 342). Gregor (1970) believed that the data were best fitted by a binomial form of an exponential decay function, with one term representing a fast decay of younger sediment and a second term representing a slower decay of older sediment. Garrels and Mackenzie (1971 a, b) interpreted the same data in terms of two cycles of sedimentation and erosion, the older ending at 350 Ma, and the younger continuing to present. Li (1972) noted that the binomial equation "is empirical and has no physical meaning" (p. 132). Dacey and Lerman (1983) found that four types of curves, geometric, exponential, compound decay, and inverse hypergeometric, all fit reasonably well to the Phanerozoic mass/age data of Gregor (1985) normalized to 50 my intervals. They used an age-dependent Markov chain model to describe the mass/age distribution as approximated by the four curves.

In spite of the fact that the binomial exponential decay equation has no unambiguous physical meaning, it has become evident that the shape

1069

1070

Christopher N. Wold and William W. Hay

of the Phanerozoic mass/age distribution is best described by two exponential decays. Using a global data set, Wold, Hay, and Wilson (1987) found that the mass/age distribution of Phanerozoic sediments could be closely approximated by the addition of two exponential decays,

y = (56.10 X 1021 g) exp [-0.002 my-1 t]

+ (73.66 X 1021 g) exp [-0.0296 my-1 tl.

They attributed the slow decay describing the distribution of ancient sedimentary rocks to the fact that the older rocks are generally more lithified and protected from erosion by burial beneath younger materials and the fast decay affecting young sediments to their unlithified condition and wide exposure. They suggested that the original flux rate of any given time interval was proportional to the ratio of the observed mass to the mass represented by the sum of the exponential decays for that interval. The analysis of the data suggested that a 150 my cycle of greater or lesser sediment flux overprints the exponential decays.

Recently, we became concerned that there may be no unique answer to the problem of reconstructing ancient sediment masses and fluxes (Wold and Hay, 1988) and suggested that there might be at least three possibilities (discussed below), depending on how the exponential decay curve is fit to the data. A least squares fit of a single exponential decay to all the data always appears to be a smooth, gentle decay that poorly represents the rapid decay of the young sediment mass: furthermore, the mass represented by the area under the curve is rarely equal to the total observed sediment mass. A forced fit of one exponential decay curve using three selected points much more closely approximates the observed sediment mass/age distribution, but there is no a priori reason for selecting any particular set of points other than, for example. to choose the set that will cause the mass represented by the area under the curve to be equal to the actual mass observed. A fit forced to pass through the youngest data point, on the assumption that it is the best known of all data points, can be made to approximate closely the young end of the curve but rarely fits the older part well.

Although the procedure for reconstructing ancient sediment fluxes (Wold, Hay, and Wilson, 1987) seemed reasonable, and it or a similar method has been used by Shaw (ms, 1988, ms), Hay, Sloan, and Wold (1988), Hay, Shaw, and Wold (1989), Hay, Wold, and Shaw (1989), Nkounkou (ms), Tardy, Nkounkou, and Probst (1989), and Wilkinson and Walker (1989) to reconstruct past fluxes in a variety of contexts, we became concerned that it might be incorrect (Wold and Hay, 1988) because: (I) there is no unique solution to the problem of fitting a curve to the data in such a way that it will appear to approximate closely the data; (2) if more than one exponential decay term is used, as suggested by Gregor (1970) and Wold, Hay, and Wilson (1987), there is no objective way to divide the sediment into that part responding to the

Estimating ancient sediment fluxes

1071

slow decay and that part responding to the fast decay: and (3) being based on exponential decay or decays with constant terms, the method for reconstruction of ancient masses and fluxes implicitly assumes that the masses of young sediment that formed at different times can be treated independently from the rates at which older sediment was eroded and might violate the principle of conservation of mass.

The purpose of this paper is to suggest that the simplest method for reconstructing ancient sediment fluxes is also the best. We then apply the method to reconstruction of Phanerozoic fluxes, using the time scale of Snelling (1985), global sediment masses given by Ronov (1980) and Budyko, Ronov, and Yanshin (1985.1987), and explore the implications of the variations in sedimentary fluxes with time.

METHOD

Although new sediment is formed from the weathering of igneous and metamorphic rocks and existing sediment is lost to subduction and metamorphism, these processes have operated during the Phanerozoic at rates that are small compared to that at which older sediment is cannibalized to form young sediment. Hence the total sediment mass can be considered to be in quasi steady state.

To simplify the calculations, we neglect the complications of specific gains and losses from the system and fluctuations of magnitude of the total sediment mass. We assume that (1) on a global scale the sedimentary system is in steady state and has a mass that remains constant; (2) most younger sediments are derived from cannibalization of older sediments through erosion: and (3) gains of sediment mass from weathering, erosion, and deposition of igneous and crystalline metamorphic rocks are exactly offset by losses to subduction and metamorphism. Then the general decline of sediment mass with age resulting from cycling of older sediment to become younger sediment may be approximated by a simple exponential decay

y = A exp [b rl.

(1)

where y is the remnant of the original sediment flux at time t, that would be observed today after t my of cycling at a constant rate of erosion b (decay constant, or "average recycling proportionality parameter" of Veizer and Jansen, 1985), and a constant depositional rate, A (the rate at which sediment is being deposited at present). Eq (1) is the average rate of sediment cycling over the time represented by the mass/age distribution being considered.

I n a more explicit geologic context, eq (1) predicts the mass of sediment of a given age that would still be in existence today if the rate of deposition and erosion (cycling) of sediment had remained constant through time. Following the terminology developed by Hay, Shaw, and Wold (1989), eq (1) can be expressed as

Msu = Ms; exp [b (t, + t)/21

(2)

1072

Christopher N. Wold and William W. Hay

where Msij is the mass of sediment of age tj to tj, predicted by the decay curve to be in existence today, and Ms~ is the average original sediment flux of an interval of length I tj - tj I centered on the present (equivalent to A to eq 1). The erosional constant b is the proportion of the sediment mass from each time interval of equal length I tj - tj I eroded to form its proportion of the sediment mass of the youngest time interval. The sediment mass of any interval is approximated by the exponential decay curve at the midpoint of that interval, (t, + t)/2, multiplied by the length of the interval ( I t, - tj I ). The expressions in eq (2) are illustrated in figure 1 and given for real data in tables 3 and 4.

From the assumption of a steady state system of constant sedimentary mass, it follows that variability within the data must be a reflection

A.

age-

• _ It MSij

MSjj - Msp M It S··

/ IJ

==>

c.

age -

-:

I

II

I I

I

I

I

I I I I I

i

I •

r---l MsiJ·

I I

I I

I I I I

Fig. I. Diagrammatic representation of the expressions used in eqs (2) and (6). (A) is an observed mass/age distribution where age is mcreasing toward the right and an exponential decay curve fit usmg the method of least squares. Ms~ IS the intercept of the curve with the ordinate. The three bars and section of the decay curve highlighted by the double box are shown in (B). MS;j is the height of or observed mass represented by any bar, Ms;: is the observed height of the exponential decay curve at the midpoint of an interval t; to ij. Eq (6) from the text is shown in the center of the diagram. (C) is the observed mass/age distribution from (A), and the total height of each bar indicated by dashed lines is the "reconstructed flux" or mass of young sediment deposited during an interval that would have existed at the end of the interval. The three bars highlighted by the double box are shown in (D). Mst is the "reconstructed flux" for any interval whose observed mass IS MS;j.

Estimating ancient sediment fluxes

1073

of both changes in the amount of sediment originally deposited and changes in the subsequent rates of erosion of older sediment or weathered igneous or metamorphic rock. If the sediment flux varies during each time interval, then the constants of deposition and decay, A and b (eq 1), or Ms~ and b (eq 2) must vary during each time interval. Furthermore, within a steady state system of constant mass, the deposition constant and erosion (decay) constant are interrelated.

To describe this interrelationship, we will assume that the total global sediment mass (TSM) remained constant during the 600 my preceding any time in the Phanerozoic, that is, from the present to 1200 Ma. Second, we assume that the original sediment flux at any moment in time during the Phanerozoic was derived from a rate of erosion (b) and deposition (A) that were constant during that moment in time. Then, because TSM is constant and A and b are constant at any moment in time in the Phanerozoic, A and b are related to TSM by the definite integral

(600

TSM = Jo A exp [b t1 dt,

(3)

over the length of the Phanerozoic (600 my), where time (t) is given in millions of years. The solution to eq (3) is

A

TSM = B (exp [600 b] - 1).

(4)

The interrelationship between A and b is such that with a TSM that is constant from the present to 1200 Ma (as described above), A can be calculated from b

bTSM

A ---~--~~--

- exp [600 b] - 1

or b could be solved by successive approximation of eq (4) if A were given. Thus, A and b are dependent on a constant total global sediment mass (TSM) and on each other.

Every change in the flux at every time in the past changed the erosion rate at which older materials were attacked. This means that the "constants" are constant only during a moment in time and, in reality, change continuously. The result is a complex process, for, as shown in figure 2, when the deposition rate per unit time increases, the erosion rate increases, but not only is all preexisting sediment more rapidly destroyed, the younger sediment is destroyed more rapidly than the older sediment. As a result the global geologic record represented by the modern sediment mass/age distribution is the product of innumerable changes in the rates of erosion and deposition. If the sedimentary system were closed the mass/age distribution would faithfully record all the changes that took place. However, we have assumed a system that,

(5)

1074

Christopher N. Wold and William W. Hay

0.8

A.

en I-

Z 0.6

::J

en

en 0.4

«

:e

0.2

o

100

200

300

400

500

AGE

0.8

B.

en I-

Z 0.6

::J

en

en 0.4

« :e

o

100

200

300

AGE

400

Fig. 2. The blank section on top of the second through the last hatched bars is the amount of material eroded from that bar to form its portion of the youngest mass of sediment shown as the bar farthest left on each diagram. (A) shows a faster constant erosion rate with b ~ -0.02, and (B) shows a slower constant erosion rate with b ~ -0.005.

although not closed, is a steady state system where mass is conserved. In this case too it should be possible, based on the simple reconstruction method described below, to deconvolute the complexly intertwined record and to calculate a value for the sediment flux at every time in the distribution.

To accomplish this and determine the past history of changes in erosion and sedimentation, we first fit a single exponential decay curve

Estimating ancient sediment fluxes

1075

(eq 1) using the method of least squares to the present mass/age distribution of sediments. Then, we assume that the average rate of sediment cycling represented by the exponential decay curve fit to the data has remained constant through the time represented by the mass/ age distribution.

Finally, we assume that variations in the data represent real variations in the amount of sediment accumulating during given time intervals, and that subsequent variations in the rates of erosion affect the preservation of all older sediments. From this it follows that for each time interval (t, to t1) represented in the mass/age distribution there is an observed mass ot sediment existing today (MsU) that differs from the mass predicted to be in existence today by the decay curve (Msij: see fig. 1). The ratio of MSij to Msij is proportional to the excess or deficit of existing sediment relative to the amount predicted by the exponential decay curve. The deposition constant Ms~ (the intercept of the exponential decay curve on the ordinate representing sedimentary mass in a mass/age diagram: for example, figs. 1,4, and 5) is in effect a proportionality constant representing the average sediment flux from which the flux at all previous times in the mass/age distribution will be determined. The original amount of sediment of age t, to tj (Msij) which would have existed (as a result mostly of cycling of older sediments but with some contribution from weathering of igneous and crystalline metamorphic rocks) at the end of the interval (t.), is then

Ms ..

• # IJ

MSij = Msp --# .

MSij

(6)

THE PHANEROZOIC MAssi AGE DISTRIBUTION

We have analyzed the mass/age distribution of existing CambrianPliocene sediments using the compilation of Ronov (1980) for Cambrian through Pliocene sediments on land, the compilation of Budyko, Ronov, and Yanshin (1985,1987) for the Late Jurassic through Pliocene sediments on submarine shelves, continental slopes, and the ocean floor (table 1). and for the masses of volcanogenic rocks for Cambrian through Pliocene (table 2). The sediment volumes given by Ronov (1980) have been converted to mass units (1021 g): this conversion will be reported elsewhere. Although there are other estimates for some of the sediment reservoirs (Hay. Sloan, and Wold, 1988) and other global estimates (Gregor, 1985), we chose to base the current analysis exclusively on the data of Ronov (1980) and Budyko, Ronov, and Yanshin (1985, 1987) to insure internal consistency.

For the analysis, we selected the Phanerozoic time scale of Snelling (1985), one of the most complete recent compilations, but have made some modifications. For the Cenozoic, we chose to use the epoch ages suggested by Berggren, Kent, and Flynn (1985), and the base of the Ordovician is assumed to be 513 Ma, as indicated by McKerrow, Lam-

1076

Christopher N. Wold and William W. Hay

TABLE I

The sums of Cambrian through Pliocene sediments on land (from Ronov, 1980), and Late Jurassic through Pliocene sediments on submarine shelves, continental slopes, and the ocean floor (from Budyko, Ronov, and Yanshin, 1985, 1987)

Stratigraphic Abb. Age of Base Interval Global Sediment
Intervals of Interval Length Sediment Surviving Mass
(Sediments) Mass (10" g/my)
(Ma) (my) (10" g)
PLEISTOCENE a 1.60
PUOCENE N2 5.30 3.70 55.11 14.89
MIOCENE Nl 23.70 18.40 153.47 8.34
OUGOCENE P3 36.60 12.90 78.33 6.07
EOCENE P2 57.80 21.20 115.35 5.44
PALEOCENE PI 66.40 8.60 33.89 3.94
LATE CRETACEOUS K2 95.00 28.60 191.90 6.71
EARLY CRETACEOUS Kl 135.00 40.00 202.76 5.07
LATE JURASSIC J3 152.00 17.00 83.28 4.90
MIDDLE JURASSIC J2 180.00 28.00 55.30 1.98
EARLY JURASSIC Jl 205.00 25.00 52.10 2.08
LATE TRIASSIC T3 230.00 25.00 80.60 3.22
MIDDLE TRIASSIC T2 240.00 10.00 33.30 3.33
EARLY TRIASSIC Tl 250.00 10.00 28.10 2.81
LATE PERMIAN P2 260.00 10.00 44.80 4.48
EARLY PERMIAN PI 290.00 30.00 85.50 2.85
PENNSYLVANIAN C2,3 325.00 35.00 80.20 2.29
MISSISSIPPIAN Cl 355.00 30.00 86.10 2.87
LATE DEVONIAN D3 375.00 20.00 81.30 4.07
MIDDLE DEVONIAN D2 390.00 15.00 81.50 5.43
EARLY DEVONIAN Dl 405.00 15.00 66.00 4.53
SILURIAN S 435.00 30.00 71.70 2.39
ORDOVICIAN 0 513.00 78.00 110.50 1.42
LATE CAMBRIAN C3 530.00 17.00 48.00 2.82
MIDDLE CAMBRIAN C2 550.00 20.00 54.60 2.73
EARLY CAMBRIAN ci 570.00 20.00 65.10 3.26 bert, and Cocks (1985). We assumed that the boundary between the Lower and Upper Permian lies at the top of the Kungurian Stage. Ronov (1980) did not subdivide the Ordovician or Silurian, but the data on masses of volcanogenic rocks presented by Budyko, Ronov, and Yanshin (1985,1987) use a two part subdivision of the Silurian and a three part subdivision of the Ordovician. Following the discussion of subdivisions of these Systems by Harland and others (1982), we have assumed the boundary between the Early and Late Silurian to be between the Ludlovian and Wenlockian Stages, the boundary between the Middle and Late Ordovician between the Caradocian and L1andeiloan Stages, and the boundary between the Early and Middle Ordovician between the L1anvirnian and Arenigian Stages. The ages for these boundaries, according to McKerrow, Lambert, and Cocks (1985) are 420, 454, and 470 Ma, respectively.

To make it easier to visualize the differences in masses we normalized the data to 10 my intervals, the finest resolution consistent with the data (tables 3 and 4). Because there are no data for the Pleistocene, the increments do not start with the present but at 1.6 Ma, the base of the Pleistocene on the Berggren, Kent, and Flynn (1985) and Snelling (1985) time scales. The existing mass of sediment or volcanogenic rocks in each 10 my interval was calculated by determining the existing mass per million years for each of the traditional geologic intervals recog-

Estimating ancient sediment fluxes

1077

TABLE 2
The masses of volcanogenic rocks of Cambrian through Pliocene age (from
Budyko, Ronov, and Yanshin, 1985, 1987)
Stratigraphic Abb. Age of Base Interval Global I/olcanogenies
Intervals of Interval Length Volcanogenic Surviving Mass
(Volcanogenics) Mass (1021 g/m y)
(Ma) (my) (1021 g)
PLEISTOCENE Q 1.60
PUOCENE N2 5.30 3.70 4.49 1.21
MIOCENE Nl 23.70 18.40 14.63 0.80
OUGOCENE P3 36.60 12.90 7.21 0.56
EOCENE P2 57.80 21.20 13.99 0.66
PALEOCENE PI 66.40 8.60 6.28 0.73
LATE CRETACEOUS K2 95.00 28.60 46.64 1.63
EARLY CRETACEOUS Kl 135.00 40.00 50.19 1.25
LATE JURASSIC J3 152.00 17.00 21.34 1.26
MIDDLE JURASSIC J2 180.00 28.00 5.71 0.20
EARLY JURASSIC Jl 205.00 25.00 5.57 0.22
LATE TRIASSIC T3 230.00 25.00 9.01 0.36
MIDDLE TRIASSIC T2 240.00 10.00 4.69 0.47
EARLY TRIASSIC T1 250.00 10.00 6.62 0.66
LATE PERMIAN P2 260.00 10.00 6.97 0.70
EARLY PERMIAN PI 290.00 30.00 23.74 0.79
PENNSYLVANIAN C2,3 325.00 35.00 10.95 0.31
MISSISSIPPIAN Cl 355.00 30.00 25.62 0.85
LATE DEVONIAN 03 375.00 20.00 9.58 0.48
MIDDLE DEVONIAN 02 390.00 15.00 8.54 0.57
EARLY DEVONIAN 01 405.00 15.00 9.35 0.62
LATE SILURIAN S2 420.00 15.00 5.07 0.34
EARLY SILURIAN SI 435.00 15.00 7.67 0.51
LATE ORDOVICIAN 03 454.00 19.00 7.78 0.41
MIDDLE ORDOVICIAN 02 470.00 16.00 18.34 1.15
EARLY ORDOVICIAN 01 513.00 43.00 8.68 0.20
LATE CAMBRIAN C3 530.00 20.00 5.04 0.25
MIDDLE CAMBRIAN C2 550.00 20.00 10.25 0.51
EARLY CAMBRIAN Cl 570.00 20.00 9.72 0.49 nized by Ronov (1980) or Budyko, Ronov, and Yanshin (1985,1987) and summing the mass increments from the portion of each traditional geologic interval within the 10 my interval. The mass/age distribution of all pre-Pleistocene Phanerozoic sediments is shown in figure 3: the figure graphically illustrates the procedure used to normalize the data. The sediment masses originally compiled in terms of traditional geologic intervals (Ronov, 1980: Budyko, Ronov, and Yanshin, 1985, 1987), expressed as masses per million years, are shown on the left, and the data normalized to 10 my intervals are shown on the right. The areas of the bars are proportional to the total mass per interval.

The "reconstructed flux" for each 1 0 my time interval is shown in tables 3 and 4. It is estimated at the midpoint of each interval, and it represents the mass of material deposited during the interval and still in existence at the end of the interval. Clearly, by the end of the interval some of the material deposited during the interval has already been destroyed by erosion and recycled into that deposited during the younger part of the interval. Hence, the "reconstructed flux" for an interval must be less than the instantaneous flux that would be measured at the end of the interval. The term "reconstructed flux" used here corresponds to "young sediment" as used by Wold, Hay, and Wilson (1987). Because sediment masses given by Ronov (1980) and sediment and

1078

Christopher N. Wold and William W. Hay

1.6=N2 1.6
5.3 Nl
23.7
36.6 1fL,_, 51.6
57.8
66.4 Pl
IK2
95 IKl 101.6
135
152 3 151.6
180 IJ2 ====l'
205 1J1 ===l 201.6
230 ,~
240 251.6
250 CP2 T1
~60 Ipl ====;l
290 IC2.3 ===t 301.6
325 ICl ~
355 351.6
375 03
390 401.6
405
435 Is ===::j
10 @ 451.6
5Ot6
513
530 £2 -------'t
550 551.6
570 £1 14 12 10

8

6

4 2 024

1021

9/m.y.

6

10

Fig. 3. Mass/age distribution of Phanerozoic sediment using data of Ronov (1980) and Budyko, Ronov, and Yanshin (1985, 1987). The distribution on the left shows sediment masses divided into geologic intervals (the abbreviations are defined in tables I and 2). The distribution on the right shows sediment masses normalized to 10 my time intervals. The mass for any interval is equal to its mass per million years (abscissa) multiplied by the length (in million years) of the interval.

volcanogenic rock masses given by Budyko, Ronov, and Yanshin (1985, 1987) are constrained by the resolution of geologic Periods and Epochs, we are not able to estimate ancient sediment or volcanogenic rock masses for age intervals shorter than 10 my, so that the "reconstructed flux" is of necessity a crude estimate that cannot take into account variations on time scales less than 10 my. Clearly, within the first 10 my of its existence, a body of sediment is in greatest jeopardy of being eroded and redeposited. The "reconstructed fluxes" of sediment or volcanogenic rocks given here are the masses of sediment or volcanogenic rocks less than 10 myoid that are estimated, using the midpoint of the interval, to have been in existence at the end of each 10 my interval.

Figure 4 shows the masses of sediment still in existence today that were deposited during each 10 my interval starting from 1.6 Ma. It also shows a least squares exponential decay fit to the data; the equation is y = (57.35 x 1021 g) exp[ - 0.0021 my "! t]. Above each of the hatched bars representing the amount of sediment in existence today is the amount calculated to have been eroded subsequent to deposition. The top of the bars is the "reconstructed flux"; the scale on the left side of the figure gives the sediment flux per 10 my. The data are presented in

Estimating ancient sediment fluxes

1079

120
100 -
80
C'l
N 60
0
,...
40
20
0
6.6 106.6

206.6

306.6

406.6

506.6

AGE, Ma

Fig. 4. The hatched region on the bottom is the mass/age distribution of continental, passive margin and ocean floor Phanerozoic sediments (after Ronov, 1980; and Budyko, Ronov, and Yanshin, 1985, 1987); an exponential decay curve (y = A explbtp has been fit to the data using the method of least squares, where A = 57.35 x I 021g my- , and b = -0.0021 my-I. Superimposed are the reconstructed average sediment fluxes for each 10 my interval.

numerical form in table 3. The decay constant (- 0.0021 my-I) is very close to that derived by Gregor (1985) from his data set ( - 0.0024 my-I). However, it is only about half as large as that suggested by Veizer and Jansen (1985) as their "preferred" recycling constant for the global sediment mass (- 0.0040 my-I), based on separate analyses of sediments in different tectonic settings.

It is evident from figure 4 and table 3 that there are significant variations in the amounts of sediment of different age still in existence today. In addition to the maximum represented by the youngest interval, centered on 6.6 Ma, sediment mass maxima are centered on 540, 390,240, and 90 Ma (Wold, Hay, and Wilson, 1987) suggesting a 150 my cycle. The repetitive cycle inherent in the data is robust and is evident even if significantly different data [modified passive margin data from Gregor (1985) used in Hay, Sloan, and Wold (1988); and ocean floor sediment mass in Hay, Sloan, and Wold (1988)] are substituted.

Assuming a steady state system of constant sediment mass, we used the method described in the previous section (eqs 2 and 6) to reconstruct the fluxes of sediment in the past and determine the amounts of sediment eroded subsequent to deposition. We believe that the variance in the amounts of sediments presently in existence compared with the

1080 Christopher N. Wold and William W. Hay
TABLE 3
The sediment masses from table 1 normalized to 10 my intervals from 1.6 Ma
(top of the Pliocene) to 561.6 Ma (fig. 3) and reconstructed (fig. 4); the
interval from 561.6 to 570 Ma (base of the Cambrian) is not included in this
analysis because it is shorter than 10 my. Values for the expressions from eqs
(2) and (6) are given for each 10 my interval
Midpoints Total Sediment Mass Ratio of Reconstructed
of 10 my Observed Predicted by Observed to Sediment Flux
Intervals Normalized Exponential Predicted
Sediment Mass Decay Curve Sediment Mass
(Ma) (1021g) (10"g) (1021 g/10 my)
[(~+~)/2] [Msij] [MS'jj] [MSi/Ms'jj] [Ms']
6.6 107.66 56.56 1.90 109.16
16.6 63.41 55.38 1.51 86.37
26.6 65.49 54.23 1.21 69.25
36.6 57.57 53.10 1.08 62.17
46.6 54.41 51.99 1.05 60.01
56.6 48.71 50.91 0.96 54.86
66.6 53.81 49.85 1.08 61.89
76.6 67.10 48.82 1.37 78.82
86.6 67.10 47.80 1.40 80.50
96.6 56.27 46.81 1.20 68.94
106.6 50.69 45.63 1.11 63.42
116.6 50.69 44.88 1.13 64.77
126.6 50.69 43.94 1.15 66.15
136.6 49.57 43.03 1.15 66.06
146.6 48.99 42.14 1.16 66.67
156.6 20.92 41.26 0.51 29.08
166.6 19.75 40.40 0.49 28.03
176.6 19.92 39.56 0.50 28.88
186.6 20.84 38.74 0.54 30.85
196.6 20.84 37.93 0.55 31.51
206.6 28.36 37.14 0.76 43.79
216.6 32.24 36.37 0.89 50.63
226.6 32.41 35.61 0.91 52.19
236.6 32.47 34.87 0.93 53.39
246.6 30.77 34.15 0.90 51.68
256.6 42.19 33.44 1.26 72.36
286.6 28.50 32.74 0.87 49.92
276.6 28.50 32.06 0.89 50.98
286.6 27.61 31.39 0.88 50.43
296.6 22.91 30.74 0.75 42.75
306.6 22.91 30.10 0.76 43.66
316.6 22.91 29.47 0.78 44.58
326.6 26.73 28.86 0.93 53.12
336.6 28.70 28.26 1.02 58.24
346.6 28.70 27.67 1.04 59.48
356.6 36.59 27.10 1.35 77.43
366.6 40.65 26.53 1.53 87.86
376.6 49.68 25.98 1.91 109.66
386.6 52.89 25.44 2.08 119.23
396.6 45.33 24.91 1.82 104.36
406.6 31.19 24.39 1.28 73.32
416.6 23.90 23.88 1.00 57.38
426.6 23.90 23.39 1.02 58.60
436.6 17.48 22.90 0.76 43.76
446.6 14.17 22.42 0.63 36.23
456.6 14.17 21.96 0.65 37.00
466.6 14.17 21.50 0.66 37.78
476.6 14.17 21.05 0.67 38.59
486.6 14.17 20.62 0.69 39.41
496.6 14.17 20.19 0.70 40.24
506.6 14.17 19.77 0.72 41.10
516.6 26.27 19.36 1.36 77.82
526.6 28.09 18.95 1.48 84.98
536.6 27.30 18.56 1.47 84.35
546.6 28.14 18.17 1.55 88.80
556.6 32.55 17.79 1.63 104.90 Estimating ancient sediment fluxes

1081

amounts hindcast by the exponential decay function (fig. 4) reflects real variations of up to threefold in the global rates of erosion and sedimentation through time. In reality, the sediment system is not closed, but additions come from the weathering of igneous and metamorphic rocks, and losses occur as sediment is subducted and metamorphosed. These gains and losses undoubtedly vary with time but probably correspond closely to the variance in sediment fluxes based on the assumption of a steady state system of constant sediment mass. The present rate of subduction of ocean floor sediment of 1 x 1021 g/my suggested by Hay, Sloan, and Wold (1988) is about one-sixth (17 percent) of the reconstructed sediment flux per million years for the 10 my immediately preceding the Quaternary and one-sixth of the Phanerozoic average. During the Phanerozoic, metamorphism not associated with subduction has probably been minimal and would represent a small additional mass loss. This suggests that if the sediment mass is in steady state, during the Phanerozoic the rate of cannibalization of older sediment has averaged about 83 percent, with gains from weathering of igneous and crystalline metamorphic rocks and losses to subduction and metamorphism averaging + and - 17 percent respectively. These values are within the upper limit on cannibalization set by Sm/Nd data collected during the 1970's (McCulloch and Wasserburg, 1978; Veizer and Jansen, 1979). Subsequent Sm/Nd data have been analyzed by Veizer and Jansen (1985), who concluded that at present the sedimentary system is 90 ± 5 percent cannibalistic.

Budyko, Ronov, and Yanshin (1985, 1987) gave masses for the amounts of "volcanogenic rocks" incorporated into the Earth's sedimentary shell and representing each major pre-Quaternary interval of the Phanerozoic. Although not stated explicitly, the compilation must rest largely on that of Ronov (1980), who presented the data in terms of volumes and volcanogenic rocks in the Earth's sedimentary shell. Ronov (1980) used 2.8 g/cc as the density for the "volcanogenic rocks," indicating that the bulk of the "volcanogenic rocks" must be basalts and other compact extrusives or intrusives intercalated with the sedimentary rocks and can only to a minor extent include ashes, tuffs, and other less dense volcanogenic materials.

Figure 5 and table 4 show the masses of volcanogenic rocks still in existence today that were erupted during each 10 my interval starting from 1.6 Ma. The least squares exponential decay fit to the data is y = (8.53 X 1021 g) exp[-0.0016 my- t]. Above each of the hatched bars representing the amount of volcanogenic rocks in existence today is the amount calculated to have been eroded subsequent to deposition. The top of the bars is the "reconstructed flux;" the scale on the right side of the figure gives the sediment flux per million years. The decay constant is slightly less than that for sediments; this is likely a reflection that the volcanogenic rocks are from the outset more indurated and less readily attacked by erosion.

1082
25
20
15
en
N
0
,... 10
5
0
6.6 Christopher N. Wold and William W. Hay

106.6

206.6

306.6

406.6

506.6

AGE, Ma

Fig. 5. The hatched region on the bottom is the mass/age distribution of Phanerozoic volcanics, the exponential decay curve fit to the data using the method of ~east squares has the parameters: A = 8.53 X 1021g my ", and b = -0.0016 my-I. Supenmposed are the reconstructed average volcanogenic fluxes for each 10 my interval.

The maxima of mass of volcanogenic rock are approximately coincident with those for sediment. The Cambrian peak is well defined and corresponds closely to the 540 Ma sediment peak. However, the midPaleozoic peak is much broader than the sediment peak although it is still centered on 390 Ma. The analysis suggests that large masses of volcanogenic rock accumulated during the Middle Ordovician, and that major accumulations continued until the end of the Devonian. The Middle Ordovician accumulation rates could be due to an underestimate of its length due to an error in either or both of the boundary ages, but later peaks also suggest that large masses of volcanogenic rock accumulate prior to the peaks of sediment accumulations. The peak of mass of volcanogenic rock at the Paleozoic-Mesozoic boundary occurs slightly earlier than the peak of sediment mass, being centered on about 360 Ma, but it has almost the same duration at the sediment mass peak. The Late Jurassic-Cretaceous peak corresponds very closely to the sediment mass peak, but the larger accumulations of volcanogenic rocks precede the largest accumulations of sediment. The analysis suggests that volcanic activity during the last 20 my has been slightly larger than that during the earlier Cenozoic but only about half that of the Cretaceous.

The total amount of volcanogenic rock calculated to have been destroyed by subsequent weathering and erosion offers a clue to the rate at which igneous material is converted to sediment. Our reconstructions

Estimating ancient sediment fluxes 1083
TABLE 4
The volcanogenic rock masses from table 2 normalized to 10 my intervals and
reconstructed (fig. 5). Values for the expression from eqs (2) and (6) are given
for each 10 my interval
Midpoints Total Volcanic Mass Ratio of Reconstructed
of 10 my Observed Predicted by Observed to Volcanic Flux
Intervals Normalized Exponential Predicted
Volcanic Mass Decay Curve Volcanic Mass
(Ma) (1021g) (1021g) (1021 g/10 my)
[(t,+W2] [MS'j] [Ms',j] [Ms,/Ms#'j] [MS·]
6.6 9.50 8.43 1.13 9.60
16.6 7.95 8.30 0.96 8.17
26.6 6.09 8.17 0.75 6.35
36.6 6.09 8.03 0.76 6.47
46.6 6.60 7.90 0.83 7.12
56.6 6.87 7.78 0.88 7.53
66.6 11.99 7.65 1.57 13.36
76.6 16.31 7.53 2.17 18.47
66.6 16.31 7.41 2.20 18.77
96.6 13.83 7.29 1.90 16.18
106.6 12.55 7.17 1.75 14.92
116.6 12.55 7.05 1.78 15.17
126.6 12.55 6.94 1.81 15.42
136.6 12.55 6.83 1.84 15.67
146.6 12.55 6.72 1.87 15.93
156.6 2.46 6.61 0.37 3.17
166.6 2.04 6.50 0.31 2.67
176.6 2.07 6.40 0.32 2.76
166.6 2.23 6.29 0.35 3.02
196.6 2.23 6.19 0.36 3.07
206.6 3.14 6.09 0.51 4.39
216.6 3.60 5.99 0.60 5.13
226.6 3.78 5.90 0.64 5.46
236.6 5.00 5.80 0.86 7.34
246.6 6.88 5.71 1.17 9.97
256.6 7.12 5.62 1.27 10.81
266.6 7.91 5.53 1.43 12.21
276.6 7.91 5.44 1.46 12.41
286.6 7.15 5.35 1.34 11.39
296.6 3.13 5.26 0.59 5.07
306.6 3.13 5.18 0.60 5.15
316.6 3.13 5.10 0.61 5.24
326.6 6.70 5.01 1.34 11.40
336.6 8.54 4.93 1.73 14.76
346.6 8.54 4.65 1.76 15.00
356.6 6.07 4.77 1.27 10.83
366.6 4.79 4.70 1.02 8.69
376.6 5.39 4.62 1.17 9.94
386.6 5.78 4.55 1.27 10.84
396.6 6.23 4.47 1.39 11.88
406.6 4.35 4.40 0.99 8.43
416.6 3.66 4.33 0.84 7.20
426.6 5.11 4.26 1.20 10.23
436.6 4.44 4.19 1.06 9.03
446.6 4.09 4.12 0.99 8.46
456.6 9.69 4.06 2.39 20.37
466.6 9.95 3.99 2.49 21.25
476.6 2.02 3.93 0.51 4.38
486.6 2.02 3.86 0.52 4.45
496.6 2.02 3.80 0.53 4.53
506.6 2.02 3.74 0.54 4.60
516.6 2.83 3.88 0.77 6.56
526.6 3.31 3.62 0.91 7.79
536.6 5.13 3.56 1.44 12.26
546.6 5.08 3.51 1.45 12.36
556.6 4.86 3.45 1.41 12.01 1084

Christopher N. Wold and William W. Hay

suggest that 186 x 1021 g of pre-Quaternary Phanerozoic volcanogenic rock has been destroyed since it was formed: most of this has been weathered, eroded, and deposited as sediment. This source of new material for the sediment system would supply only about one-third of the amount of sediment lost to subduction. If steady state is to be maintained, it requires that the weathering of intrusive igneous and crystalline metamorphic rocks supply twice as much material to the sedimentary system. Blatt and Jones (1975) estimated that extrusive igneous rocks cover 8 percen t of the Earth's land surface, igneous intrusive rocks 9 percent, and metamorphic and "Precambrian" (which is mostly metamorphic, but may include metasediments and sediments) rocks 17 percent. The percentage of areal exposure on land suggests that over the long term extrusive igneous rocks, intrusive igneous rocks, and crystalline metamorphic rocks may weather in roughly equal proportions.

Figure 6 shows the proportion of volcanogenic rocks as a part of the total of sediments plus volcanogenic rocks. Four major peaks are discernible, with a minor peak in the Cambrian. The Middle Ordovician appears to be a much larger peak than the others, and this cannot be due simply to an error in the estimate of its length; even taking the Ordovi-

VI
c
Z 0.5
W
e
0
z
<
c 0.4
..J
0
>
Q
Z
<
VI 0.3
I-
Z
w
:::E
2i
W 0.2
VI
..J
<
I-
0
l-
LL 0.1
0 ~
z
0
f=
a:
0 0
0. 6.6 106.6 206.6 306.6 406.6 506.6
0
a:
0. AGE, Ma Fig. 6. The proportion of existing volcanogenic as a part of the total of sediments plus vofcanogenic sediments for the Phanerozoic. The almost straight line, an exponential curve y = 0.147 exp[0.000023 t ], fit to the data indicates that the long term average proportion of volcanogenic to total sediment has been almost constant through the Phanerozoic and is 15 percent.

Estimating ancient sediment fluxes

1085

cian as a whole, the ratio of volcanogenic rocks to sediment is unusually high. The Devonian peak and that at the end of the Paleozoic are sharp and well defined. The Late Jurassic-Cretaceous peak is broad and corresponds to the peaks for volcanogenic rocks and sediments.

A regression line fitted through the proportion of volcanogenic rocks is almost flat, declining very slightly and probably insignificantly from the beginning of the Cambrian to the Pliocene. This suggests that although there are cyclic variations in volcanic activity, no long term decline can be detected in the data for the Phanerozoic.

IMPLICATIONS

In seeking the underlying causes of the fluctuations of volcanism, erosion, and sedimentation, both endogene and exogene causes must be considered. The volcanic fluxes are directly related to endogene dynamics but could reflect three possible types of volcanism: (1) activity at constructive plate margins, that is, spreading centers such as mid-ocean ridges and rifts: (2) activity at destructive plate margins, that is, subduction zones; and (3) intraplate volcanism, usually attributed to hot spots and upwelling plumes from deep within the mantle. In Ronov's (1980) summary compilation volcanogenic materials are reported as terrestrial or submarine and are divided between platform and geosynclinal regions. In the volcanic masses reported for the later Mesozoic and Cenozoic by Budyko, Ronov, and Yanshin (1985, 1987) totals are given for continental blocks (platforms + geosynclines), submarine shelves and slopes, and for the ocean floor (not including the ocean crust proper). The largest categories are submarine volcanogenic rocks from geosynclinal regions and from submarine shelves and slopes. It seems evident that intraplate volcanism is a relatively small contribution, and constructive plate volcanism is also a small contributor to the volcanogenic mass. The dominant mass of volcanogenic rocks (appro x 80 percent of the total) is subduction related, ultimately being included in "geosynclinal deposits." As a working hypothesis for geochemical modelling, changes in the rate of sea-floor spreading have been interpreted as directly proportional to changes in volcanic activity (Berner, Lasaga, and Garrels, 1983). It is therefore reasonable as a working hypothesis to assume the converse: that changes in the rate of production of volcanic materials are an index to ancient rates of sea-floor spreading, and that the latter is a direct measure of the rate of plate tectonic processes generally.

Of the total sediment, about 80 percent is detrital: the changes in sediment flux through time must then be primarily due to changes in the flux of detrital sediment. Most detrital sediment is transported by rivers from its site of origin, and the suspended load of rivers is foremost a function of the elevation of the drainage basin (Hay and others, 1987:

Pinet and Souriau, 1988). Hence, we believe that the globally varying rates of sediment fluxes are primarily a reflection of the rates of uplift and mountain building. The synchroniety of changes in volcanic and

1086

Christopher N. Wold and William W. Hay

sediment fluxes (figs. 4 and 5) suggests that global rates of volcanism, uplift, and mountain building are coupled responses to changes in the rates of plate tectonic activity.

The variations of surficial sediment fluxes indicated in figure 4 imply changes by a factor of three or more in surficial geochemical cycles.

It is usually thought that rapid sea-floor spreading produces a mid-ocean ridge system that displaces a larger volume of water, causing the ocean to flood the continents: however, only two of the three sediment and volcanic flux maxima, that in the middle Paleozoic and that in the late Mesozoic, correspond to times of continental flooding. The late Paleozoic-early Mesozoic flux maximum corresponds to a relatively low stand of sealevel. However, there is abundant evidence for high rates of plate tectonic activity in the late Paleozoic and early Mesozoic as small continental fragments were repeatedly split off Gondwanaland and rapidly crossed the ancestral Tethys to collide with nuclear Asia a few tens of millions of years later (Sengor, 1987: Wilson, Rosol, and Hay, 1989). This style of late Paleozoic-early Mesozoic plate tectonics was fundamentally different from that in the Jurassic and Cretaceous when the great continental blocks of Laurasia and Gondwana ruptured through the middle to produce the present continents that are still adrift. In contrast to the Cretaceous episode of rapid sea-floor spreading, when the mid-ocean ridge displaced large volumes of water, sea-floor spreading in the late Paleozoic-early Mesozoic involved a ridge system that was largely marginal to the ocean and repeatedly jumped back under the margin of Gondwanaland to split off fragments of the continent. Because the ridge was associated with the continental margin, it did not displace large volumes of sea water and did not result in flooding of the continent.

The effect of changes in the rates of plate tectonic processes on climate are threefold: (1) changes in the composition of the atmosphere resulting in changes in the radiation balance (probable high levels of CO2 as a result of rapid sea-floor spreading in the Cretaceous as a cause of warm polar temperatures): (2) changes in continental relief resulting in changes in atmospheric circulation patterns: and (3) changes in the areas of shallow seas (as a result of continental flooding) causing changes in albedo and/or cloud cover. The changes in climate may affect erosion rates by changing the nature and rates of weathering and preventing soil formation from reaching equilibrium. An implication of the changing sediment flux rates is that the climate system is rarely stable for long periods of time.

SUMMARY AND CONCLUSIONS

Realization that the mass/age distribution of sediments reflects a cannibalistic cycling process has come only in the last two decades. Because of recycling, the present mass/age distribution of sediments is not a direct reflection of the original flux rates but contains information

Estimating ancient sediment fluxes

1087

permIttIng the original flux rates to be reconstructed. The simplest method of estimating the original sediment flux rates is to fit an exponential decay curve to the observed mass/age distribution and then multiply the ratio of the observed mass to the mass indicated by the decay curve of the O-age flux rate predicted by the exponential decay curve.

Applying the reconstruction method to the mass/age distribution of existing Phanerozoic sediments and volcanics indicates that for both these materials the original global flux rates have varied by a factor of three. The variations of flux rates of volcanogenic rocks and sediments are synchronous. The reconstructed flux of Phanerozoic volcanogenic rocks is considered to be a direct index to changing rates of plate tectonic activity. The reconstructed flux of Phanerozoic sediments is considered to be a reflection of changes in the area and average elevation of land, implying significant changes in the rates of uplift and mountain building reflecting global plate tectonic activity.

ACKNOWLEDGMENTS

The authors have benefitted from conversations with a number of colleagues: Robert M. Garrels, Fred T. Mackenzie, C. Bryan Gregor, Robert A. Berner, Heinrich D. Holland, John R. Southam, Lee Kump, Yves Tardy, and Renard-Roger Nkounkou. This work has been supported by grants OCE 8409369 and OCE 8716408 from the National Science Foundation, grants 19274-AC2 and 21626-AC8 from the American Chemical Society'S Petroleum Research Fund, and by gifts from Texaco, Inc. The authors are grateful for the generous use of the facilities of both the Institut fiir Allgemeine und Angewandte Geophysik and the Institut fur Palaontologie und Historische Geologie of Ludwig-Maximillians Universitat, Munich, kindly made available through Professors Heinrich Soffel and Dieter Herrn, and especially to Dr. Peter Neurieder for his patient assistance.

REFERENCES

Berggren, W. A., Kent, D. V., and Flynn, J. J., 1985, Jurassic to Paleogene: Part 2 Paleogene geochronology and chronostratigraphy, in Snelling, N.J., ed., The Chronology of the Geological Record: Oxford, Blackwell Scientific Publications, Geological Society (London) Memoir 10, p. 141-195.

Berner, R. A., Lasaga, A. C., and Garrels, R. M., 1983, The carbonate-silicate geochemical cycle and its effect on atmospheric carbon dioxide over the past 100 million years:

AmericanJournal of Science, v. 283, p. 641-683.

Blatt, H., and Jones, R. L., 1975, Proportions of exposed igneous, metamorphic and sedimentary rocks: Geological Society of America Bulletin, v. 86, p. 1085-1088.

Budyko, M. I., Ronov, A. B., and Yanshin, A. L., 1985, Istoriya Atmosferi: Leningrad, Gidrometeroizdat.

---- 1987, History of the Earth's Atmosphere: New York, Springer-Verlag, 139 p. Dacey, M. F., and Lerman, A., 1983, Sediment growth and aging as Markov chains:

Journal of Geology, v. 91, P: 573-590.

Garrels, R. M., and Mackenzie, F. T., 1971a, Gregor's denudation of the continents:

Nature, v. 231, p. 382-383.

---- 1971b, Evolution of Sedimentary Rocks: New York, Norton, 397 p.

Gilluly, J., 1969, Geological perspectives and the completeness of the geologic record:

Geological Society of American Bulletin, v. 80, p. 2303-2312.

1088

Christopher N. Wold and William W. Hay

Gregor, C. B., 1970, Denudation of the continents: Nature, v. 228, p. 273-275.

1985, The mass-age distribution of Phanerozoic sediments, in Snelling, N.J., ed., The Chronology of the Geological Record: Oxford, Blackwell Scientific Publications, Geological Society (London) Memoir 10, p. 284-289.

Harland, W. B., Cox, A. V., Llewellyn, P. G., Pickton, C. A. G., Smith, A. G., and Walters, R., 1982, A Geologic Time Scale: Cambridge, Cambridge University Press, 128 P:

Hay, W. W., Rosol, M. J., Sloan, J. L. II, and Jory, D. E., 1987, Plate tectonic control of global patterns of detrital and carbonate sedimentation, in Doyle, L. J., and Roberts, H. H., eds., Carbonate-Clastic Transitions: Developments in Sedimentology: Amsterdam, Elsevier, v. 42, p. 1-34.

Hay, W. W., Shaw, C. A., and Wold, C. N., 1989, Mass-balanced paleogeographic reconstructions: Geologische Rundschau, v. 78, p. 207-242.

Hay, W. W., Sloan, J. L., II, and Wold, C. N., 1988, Massi Age distribution and composition of sediments on the ocean floor and the global rate of sediment subduction:

Journal of Geophysical Research, v. 93, p. 14933-14940.

Hay, W. W., Wold, C. N., and Shaw, C. A., 1989, Mass-balanced paleogeographic maps:

Background and input requirements, in Cross, T. A., ed., Quantitative Dynamic Stratigraphy: Englewood Cliffs, N .1., Prentice-Hall, Inc., p. 261-275.

Li, Y. H., 1972, Geochemical mass balance among lithosphere, hydrosphere, and atmosphere: AmericanJournal of Science, v. 272, p. 119-137.

McCulloch, M. T., and Wasserburg, G. J., 1978, Sm-Nd and Rb-Sr chronology of continental crust formation: Science, v. 200, p. 151-187.

McKerrow, W. S., Lambert, R. St. I., and Cocks, L. R. M., 1985, The Ordovician, Silurian and Devonian periods, in Snelling, N.J., ed., Chronology of the Geological Record:

Oxford, Blackwell Scientific Publications, Geological Society (London) Memoir 10, p.73-80.

Nkounkou, R.-R., ms, 1989, Hydrogeodynamique actuelle du Congo et de I'Amazone.

Cycle global de I'eau et bilan de l'erosion au cours des temps Phanerozoiques (derniers 600 millions d'annees): These, U niversite Louis Pasteur, Strasbourg, France, 183 p.

Pi net, P., and Souriau, M., 1988, Continental erosion and large-scale relief: Tectonics, v. 7, p. 563-582.

Ronov, A. B., 1980, The earth's sedimentary shell (quantitative patterns of its structure, compositions, and evolution).- The 20th V. 1. Vernadski Lecture, March 12, 1978 (in Russian), in Yaroshevskii, A. A., ed., The earth's sedimentary shell (quantitative patterns of its structure, compositions, and evolution): Moscow, Nauka, 80 p. [English translation in International Geology Review, v. 24, p. 1313-1388 (1982); also American Geological Institute Reprint Series, v. 5, 73 p. (1983)].

Ronov, A. B., and Yaroshevskiy, A. A., 1969, Chemical composition of the Earth's crust:

American Geophysical Union Transactions, v. 13, p. 37-57.

Senger, A. M. C., 1987, Tectonics of the Tethysides, Orogenic collage development in a collisional setting: Annual Reviews of Earth and Planetary Science, v. 15, p. 213-244.

Shaw, C. A., ms, 1987, Balanced paleogeographic reconstructions of the northwestern Gulf of Mexico marB"in and its western-central North American source area since 65 Ma: M. S. thesis, University of Colorado, Boulder, 285 p.

1988, Balanced paleogeographic reconstructions of the northwestern Gulf of Mexico margin and its western-central North American source area since 65 Ma: U.S. Geological Survey Open-File Report 88-683, 227 p.

ms, 1989, Mass balanced paleogeographic modeling: examples from the western North Atlantic Ocean and Gulf of Mexico: Ph.D. thesis, University of Colorado, 381 p.

Snelling, N. J., 1985, An interim time scale, in Snelling, N. J., ed., Chronology of the Geological Record: Oxford, Blackwell Scientific Publications, Geological Society (London) Memoir 10, p. 3-9.

Tardy, Y., Nkounkou, R.-R., and Probst,J.-L., 1989, The global water cycle and continental erosion during Phanerozoic time (570 my): American Journal of Science, v. 289, p.455-483.

Veizer,J., 1988, The evolving exogenic cycle, in Gregor, C. B., Garrels, R. M., Mackenzie, F. T., and Maynard, J. B., eds., Chemical Cycles in the Evolution of the Earth: New York, WileJ-Interscience, p. 175-220.

Veizer, I., and ansen, S. L., 1979, Basement and sedimentary recycling and continental evolution: Journal of Geology, v. 87, p. 341-370.

1985, Basement and sedimentary cycling-2: Time dimension to global tectonics: Journal of Geology, v. 93, p. 625-664.

Estimating ancient sediment fluxes

1089

Wilkinson, B. H., and Walker,j. G., 1989, Phanerozoic cycling of sedimentary carbonate:

American journal of Science, v. 289, p. 525-548.

Wilson, K. W., Rosol, M. I-. and Hay, W. W., 1989, Global Mesozoic reconstructions using revised continental data and terrane histories: A progress report, in Hillhouse,j. W., ed., Deep Structure and Past Kinematics of Accreted Terranes: Washington, DC, American Geophysical Union, American Geophysical Union/International Union of Geodesy and Geophysics, Monograph 5, p. 1-40.

Wold, C. N., and Hay, W. W., 1989, Estimating ancient sediment masses and flux rates:

Geological Society of America Abstracts with Program, v. 20, p. A 78.

Wold, C. N., Hay, W. W., and Wilson, K. M., 1987, Reconstructed mass-age distributions of young sediment through the Phanerozoic: Geological Society of America Abstracts with Program, v. 19, p. 895.

REVIEW

The Origin of Species Revisited: The Theories of Evolution and of Abrupt Appearance (2 v.); by W. R. Bird. New York, 1989 (Philosophical Library, $65.00)-Shortly after I published an essay on Darwin's Three Mistakes in Geology, I received a thick package of manuscript and a letter from Mr. W. R. Bird. He asked me to make comments on volume 1 of his opus, which was aimed at advancing the creed of "Creation science." I was probably one of the many contacted by Mr. Bird, and he was able to obtain favorable comments from several "evolutionists" to be cited on the back cover of his two volumes. Mr. Bird is a lawyer, and he argued the case on this issue before the U.S. Supreme Court. I was chosen, because Mr. Bird recognized my anti-Darwinian pronouncement. Assuming two alternatives only, all scientists who presented "data that involved weaknesses of the theory of evolution" (p. 1-475) were cited as supporters of the alternative "Theory of Abrupt Appearance." I looked over the manuscript and found it apparently a revision of a brief Mr. Bird may have prepared for his appearance before the Supreme Court. Not wishing to get involved in an issue with which I am not familiar, I did not write the comments requested by him. After the book was published, I was sent the two volumes by the publisher with a request to write a review for their newsletter or for a journal which I could select. I hesitated but decided that a review would be useful in order to clarify the muddied situation created by a very clever author.

Volume I has been advertised as an assessment of "scientific claims of the two chief views concerning the origins of the universe, the first life, and plants and animals." It is not a scientific assessment. As Gareth Nelson of the American Museum of Natural History wrote, Mr. Bird's criticism of evolutionary theory "is loaded like the proverbial pair of dice. Indeed, when Mr. Bird rolls for evolutionary theory, who would expect anything but snake eyes to come up?" One cannot expect a defense lawyer to discuss evidence detrimental to the case of his defendant; we scientists do things differently. Science progresses through falsification: Unfavorable evidence has to be considered, discussed, and the theory is improved when it is modified to such an extent that it cannot be falsified by the originally disturbing data. A lawyer argues for a case, whereas scientists search for truth: they have no defendants.

What is good legal practice is often bad science. A legal document could be written in such a way that it has a flexibility in order to respond to any unforeseeable circumstances. Good science, of course, is a synonym of clarity. Mr. Bird's opus argues for "The Theory of Abrupt Appearance," but I cannot find out what this theory is. In his definition, Bird stated: "The Theory of Abrupt Appearance is defined as scientific interpretation of scientific data postulating origin through discontinuous abrupt appearance in complex form." What is meant by "scientific?" By "discontinuous?" By "abrupt?" By "appearance?" Bird did discuss the definition of science in volume II by criticizing existing definitions, especially the one given by the McLean opinion of a 1982 trial (p, 1-20). He took the so-called anti-definitional approach and stated that science

1090