The Paleoproterozoic Era

The Paleoproterozoic Era of the Proterozoic Eon: 2500 to 1600 million years ago

Hadean Archean    Eoarchean    Paleoarchean    Mesoarchean    Neoarchean Proterozoic    Paleoproterozoic       Siderian         Rhyacian         Orosirian         Statherian      Mesoproterozoic    Neoproterozoic Phanerozoic

Introduction References

Introduction

The Paleoproterozoic is, by quite a wide margin, the longest era of geologic time. It covers 900 million years, about 20% of the entire history of the Earth. This era saw the evolution of most types of bacteria with which we are familiar today, and the earliest eukaryotes. Most importantly, the early Paleoproterozoic included the Great Oxygenation Event ("GOE") during which, for the first time, atmospheric free oxygen exceeded 0.001% of its present atmospheric level ("PAL"). ATW090224.

Geography

Modern Plate tectonics began with the Paleoproterozoic. The Paleoproterozoic was the era of continental shield formation. By and large, the Earth's Archean crust seems to have been both fragmented and somewhat unstable. Some paleogeographers assert that an episode of continent formation -- in fact a supercontinent -- was present at the end of the Archean. Kump & Barley (2007). However, if that was the case, then those continents were unstable and disappeared without a trace over the next few hundred My. The majority view is that modern style continents and familiar plate tectonics began not long before the Paleoproterozoic.

Continental shields formed from small cratons. It was during the Paleoproterozoic that small islands of crust were first stitched together to form the stable nuclei of the continents we know today. This may something of an overstatement, since relatively broad islands of Archean stability are found in the rocks northeastern Canada and Greenland (the Laurentian or Canadian Shield), Western Australia (Pilbarra Craton), and South Africa (Kapvaal Craton). These became the nuclei of the North American, Australian, and (in part) African continents, respectively. However, even in these cases, the continental craton in its present form was the product of suturing several smaller units. That suturing process largely occurred in the Paleoproterozoic. In other cases (e.g., India, South America, and North China), both crust and shield were largely products of the Paleoproterozoic.

Now that we have extruded this patently over-broad generalization, we had best defend the thesis with some concrete examples.

For example, the core of South America formed around Amazonia in the Paleoproterozoic. The geologically stable core of South America is the Amazonian craton, roughly coterminous with northern and central Brazil and the inland areas of Venezuela, both Guyanas, and Suriname. Most of western South America is composed of ephemeral orogenic mountain ranges which come and go on timescales of a few 100 My. Other bits and pieces have joined (Uruguay) or left (Central Texas?) Amazonia at various times in the geological past. However, the unchanging hub of all this activity was Amazonia. The only other significant cratons now associated with South America, the São Francisco and Rio de la Plata, are both immigrants from Africa. Iacumin et al. (2001).

The only large stretches of Archean basement remaining in Amazonia are located in the eastern section of Amazonia, mostly in the southeastern corner. Most of the rest of Amazonia was intruded and sutured together in the Paleoproterozoic. The only significant exception is the northwestern Rio Negro Province, which lies along the Brazilian-Columbian border. This province formed as an extension of Amazonia in the Mesoproterozoic. Tassinari & Macambira (1999); Sial et al. (1999). For the subsequent development of the region, see Brito Neves et al. (1999).

Baltica, the core of Europe, formed from the merger of three cratons in the Paleoproterozoic. The formation of Baltica - the continent which was to become Europe -- is one of the best-known examples. Baltica formed in the Paleoproterozoic from the fusion of three cratons: Fennoscandia (Scandinavia, the Baltics, Belarus, Eastern Poland, part of Scotland, and northern European Russia), Volgo-Uralia (the Volga Basin of Russia), and Sarmatia (the Trans-Caucasus region, the Ukraine, Moldavia and part of Romania). Gee & Stephenson (2006). The process of consolidation was complete by the end of the Paleoproterozoic. Virtually all further growth in the Proterozoic came by way of extensional tectonics and the incorporation of bits and pieces of other adjacent continents. Bingen et al. (2008).

India has a similar history. Similarly, India appears to be an amalgamation of four cratons. One of these is a small, late accretion to the southern tip (southern Tamil Nadu and Kerala). The rest consists of three Archean cratons which consolidated at the end of the Paleoproterozoic. Sankaran (1999).

The shift in continent-building style is correlated with a shift in large volcanic belts from marine to terrestrial settings. Recently Kump & Barley (2007), devised an ingenious test of the general concept. They collected a large database of reasonably characterized "large igneous provinces." LIPs are broad areas of volcanic activity. They are usually manifestations of the chafing and irritation which occurs when two cratons come in contact. During the Archean, the vast majority (80% or more) of LIPs happened under water. At the beginning of the Proterozoic, the proportions abruptly reverse. About 80% of known Proterozoic LIPs were terrestrial. The most parsimonious explanation is that cratons were now consolidating, so that the boundaries between adjacent cratons most often lay in the interior of larger masses -- continents.

The trigger may have been the accumulation of a critical amount of rigid continental crust. In fact, something more fundamental may have happened -- a change in the tectonic behavior of cratons somewhat analogous to a change of state between two crystal forms. The break between Archean and Proterozoic LIP locations is quite sharp, and the ~80% level is fairly steady for the rest of Earth history. The Early Paleoproterozoic is also the earliest time that normal plate boundaries, boundaries between essentially rigid crust elements, are seen in the geological record. Stanley (1998). Stanley also notes that the total volume of continental crust first approached present value at the end of the Archean. It seems likely that the volume of continental crust, the formation of continental shields, and the development of "normal" plate tectonics are related, although the mechanics have not been worked out. ATW090224.

Geology and Geochemistry

Introduction to the Jargon

The Great Oxygenation Event: oxygen in the atmosphere and formation of an ozone layer which blocked ultraviolet radiation. The most significant event of the Paleoproterozoic is surely the Great Oxygenation Event. Shortly after the beginning of the Paleoproterozoic, something drastic happened (or began to happen) to the atmosphere. Oxygen increased from being an insignificant trace gas to at least 0.001% of its present level, and probably to between 1% and 10% of the present atmospheric level. Even the 0.001% number would cause the formation of an ozone layer, which would block lethal short wavelength ultraviolet radiation from reaching the surface. It would also supply enough oxygen to the surface waters to permit the evolution of bacteria with the ability to live by using oxygen to digest organic compounds. Presumably, all this was due to the evolution of oxygen-producing cyanobacteria.

Most of the data is geochemical. Most of what we actually know about the Paleoproterozoic is geochemistry. This, admittedly, is dull stuff. A couple of paragraphs might have been enough to stupefy your average ravening horde of Mongol invaders and their horses. We are also bitterly and resentfully aware that Andrew H. Knoll of Harvard is the only living organism with the ability to write clearly on this subject. This is monstrously unfair. Worse, much as we would like to pass the buck ‒ even to some Yankee wingnut who wears two wristwatches ‒ it has simply become too important for us to ignore.

An example of geochemical jargon. Geochemistry tends to be written in an alien jargon, rich in subscripts and allusions to the inorganic chemistry we all forgot after freshman year. Consider the following from Bekker et al. (2004):

Sedimentary successions of age ≥2.45Gyr include placer deposits that contain detrital uraninite, siderite and pyrite, reduced shallow-water facies of iron formations, highly carbonaceous shales that are not enriched in redox-sensitive elements and palaeosols that are not oxidized.

Early diagenetic pyrite in these successions has δ³⁴S values consistent with a seawater sulphate content of <200 µM ... . In contrast, sedimentary successions younger than 2.22Gyr contain red beds, CaSO₄-rich evaporites, and shallow-water iron formations that are oxidized. These successions overlie oxidized palaeosols and have δ³⁴S records that are consistent with seawater sulphate concentrations >200 µM.

... The recent discovery of mass independent fractionation (MIF) in sulphur isotopes has provided a new tool for tracing changes in the oxygen content of the atmosphere. Sulphur of sulphides and sulphates from sedimentary units older than 2.47 Gyr has values of MIF (expressed in terms of Δ³³S) ranging from -2.5‰ to +8.1‰ ... . The only known mechanism for producing MIF in sulphur isotopes is photodissociation in the gas phase ... . Preservation of large MIF signals in the Archean record is probably related to the lack of an ozone shield in the atmosphere, allowing deep penetration of high energy ultraviolet and photochemical dissociation of SO₂ into elemental and water-soluble S species. ATW090224.

Oxidation State (of Sulfur and Other Things)

An increase in oxidation level of many elements by 2200 Mya. The first point is not so hard. The authors are talking about previous work on compounds (uranium, sulfur, iron) which have more than one oxidation state. Before the Paleoproterozoic and in the earliest Siderian (i.e. ≥2450 Mya), these elements tend to be found in reduced form. After the Early Rhyacian (i.e. younger than 2220 Mya), these elements are more often found in their oxidized forms. The most natural explanation for the increase in oxidation levels is an increase in oxygen. Canfield (2005). The implication is that that oxygen became more common in the Paleoproterozoic than it had been in the Archean.

Oxygen did not reach the deep ocean. It is important to keep in mind, however, that this oxygenation was probably restricted to the atmosphere and the surface waters. In the depths, and even in the relatively shallow waters of the outer continental shelves, the oceans remained anoxic. At least most workers seem to think so. There are some possible dissenters (e.g. Rouxel et al., 2005). Sadly, we confess to being rather vague on why the depths remained anoxic throughout the entire Proterozoic. Since some ocean mixing was bound to occur, considerable oxygen must surely have reached ocean bottoms at some point in the roughly 1.6 billion years between the GOE and the end of the Proterozoic. It follows that something was removing this oxygen with high efficiency. The oxygen sink is thought to be iron, but the mechanism eludes us.

The deep oceans were anoxic and sulfidic. Poulton et al. (2004) note that banded iron formations peter out at the base of the Statherian Period (about 1800 Mya), and that the deep ocean becomes sulfidic about the same time. See also Anbar & Knoll (2002). They favor a mechanism in which small amounts of atmospheric oxygen caused weathering of sulfur as sulphates (SO₄^-2). This was washed into the ocean, and reacted with dissolved iron, re-reducing the sulfur and oxidizing the iron. Both were then deposited as insoluble pyrite (= iron sulfide, FeS). But things are probably more complicated, and there are many opinions about the particulars of the process (e.g., Canfield (2005); Anbar & Knoll (2005)). It seems likely that the process, whatever it was, was mediated by bacteria. Anbar & Knoll (2002); Rouxel et al. (2005). A remarkable variety of bugs are capable of using (and thus reducing) ferrous iron and other metals to oxidize organic compounds. Kieft et al. (1999). However, if we are to understand why banded iron deposition was replaced by sulfidic waters and pyrite iron deposition after 1800 Mya, we'd have to have a good handle on how and why banded iron formed in the first place. But banded iron formations remain very poorly understood. See Kappler et al. (2005) for one (randomly chosen) recent hypothesis.

Molybdenum as a proxy for continental weathering; suggests that there was also a brief, latest-Archean oxygenation. Most recently, two groups have looked at ancient molybdenum (Mo). Molybdenum has several useful properties. Most notably, it is introduced into seawater almost entirely by oxidative continental weathering and removed from seawater only in euxinic environments. Both groups therefore attempted to use molybdenum measurements from various Proterozoic black shales (presumably produced in euxinic environments) as a proxy for oxidative weathering. Anbar et al. (2007) examined latest Neoarchean shales in Australia. Their findings suggest that a minor oxidation event, the "Whiff" episode likely occurred even before the Paleoproterozoic. Two days before this paragraph was written (090222) an important confirming paper was published, which we have not yet seen.

The Paleoproterozoic history of molybdenum may indicate a slow, late rise in oxygen, rather than a sudden GOE. The results from the second group, Scott et al. (2008) are summarized in the image. Note the very broad, low peak extending from the latest Archean and continuing into the Mesoproterozoic. The authors argue that the late Paleoproterozoic decline in Mo is due to the spread of deep-water sulfidic conditions. That is, the later Paleoproterozoic oceans had a great deal of euxinic water which depleted the available molybdenum. On the other hand, Scott et al. don't really explain why Mo mobilization peaks so long (200-300 My) after the conventional date of the GOE, as determined from sulfur MIF (discussed below). We suspect that, like the banded iron story, this is an indication that the actual rise of atmospheric oxygen was an extremely slow process. The GOE, as measured by the usual isotopic proxies, may represent only an early stage -- and not even the earliest stage -- of a development which continued throughout the Paleoproterozoic. For the time being, just keep that suspicion in mind. After much more tedious isotopic analysis, we'll develop the thought a bit more in the Bacteria section. ATW090224.

Sulfur Geochemistry

Sulfur Isotope Fractionation

Some reactions tend to separate ("fractionate") naturally-occurring isotopes. But, returning to our quotation from Bekker et al. (2004), what is this "δ³⁴S" stuff? Sulfur has four stable isotopes: ³²S, ³³S, ³⁴S, and ³⁶S. The light (³²S) isotope, has 16 protons and 16 neutrons in its nucleus. The other isotopes differ in having more neutrons. ³²S accounts for 95% of all sulfur atoms on Earth. Most of the rest is ³⁴S. Thus, when looking at isotope ratios, it is often convenient to look at ³⁴S/³²S, compared to the same ratio in a standard material. That comparison (skipping a lot of annoying computation) is expressed as δ³⁴S [4]. The δ³⁴S differs among materials in nature because some naturally occurring reactions run slightly faster with one isotope than with another. A reaction which favors one isotope over another, leaving a product with an altered isotope ratio, is said to fractionate the element.

An example of a biological mechanism which fractionates sulfur. For example, some bacteria fractionate sulfur by converting sulfur or sulfides (e.g. hydrogen sulfide gas, H₂S) into sulfates. Compare the behavior of two hydrogen sulfide molecules, one with ³²S (light), and the other with ³⁴S (heavy) sulfur. Recall that the mass of a molecule is virtually all in the nuclei, while size, shape and chemistry is virtually all determined by the outer electron shell. The light molecule, having 2 fewer neutrons, is ~6% lighter than the heavy molecule. But molecules of H₂S have the same size, regardless of the sulfur isotope, since their electron clouds are the same; and the outer electron cloud determines size, shape, and chemical reactivity. Now, at the molecular level, temperature reflects the average kinetic energy ( = mv²/2) of the molecules in the medium. Thus, on average, the lighter molecules move faster (by 2.5% in this case) at any given temperature. Since the light and heavy molecules are the same size and shape, the faster-moving light molecules, sweep out a larger volume (7.7% larger in this case) in any given time interval. Consequently, the light hydrogen sulfide molecules are more likely to run into a bacterial enzyme and be converted to sulfate. Over time, we may find that sulphates are relatively enriched in ³²S ("light"), while sulfides are relatively enriched in ³⁴S ("heavy") sulfur.

Fractionation results reflect competing reactions and may be hard to interpret. However, life isn't that simple. There are other bacteria which perform the reverse reaction and reduce sulfates to sulfides. Non-biological reactions may also produce, bury, oxidize, or reduce sulfur. All of these processes fractionate sulfur to one degree or another. In fact, generally speaking, δ³⁴S tends to be high when the environment has lots of sulfate, oxygen, and biological activity. Sulfates from the Rhyacian and later have a significantly higher δ³⁴S than Archean sulfur. This is useful, but a little vague, since lots of different factors can fractionate sulfur (that is, enrich one isotope with respect to another). ATW090224.

Mass-Independent Fractionation

A very few reactions fractionate sulfur isotopes based on factors other than atomic mass. Thus, discussions of sulfur isotope fractionation in the Proterozoic became highly technical, complex, and unsatisfactory by about 2000. See review by Canfield & Raiswell (1999). However, a few reactions exhibit "mass-independent fractionation." These reactions tend to react preferentially with different isotopes based on nuclear spin states, thermal neutron cross-section, and ... but who are we kidding? You have no idea what we're talking about, and neither do we. The point is that these are subtle quantum properties which do not relate to mass in the simple way explained above.

These mass-independent fractionation (MIF) reactions are caused by ultraviolet radiation. Significantly, the only reactions likely to cause mass-independent fractionation under natural conditions are high-energy photochemical reactions. For these reactions to occur at all, highly energetic photons (that is, short wavelength ultraviolet light) must penetrate deep into the atmosphere to react with atmospheric sulfur near the surface. In an atmosphere with more than trace amounts of oxygen, atmospheric sulfur is rapidly (by geological standards) oxidized to sulfate and removed to surface water as acid rain. Furthermore, high energy UV never gets low enough to react with this sulfur, since it is absorbed by an ozone layer.

MIF, and thus UV radiation, can be determined separately from other fractionation. Thus the very existence of mass independent fractionation would tell us a lot -- if we could separate it out from other types of isotope fractionation. We can do just that, because sulfur has multiple stable isotopes. Imagine some environmental mass-dependent fractionation which favors ³³S over ³²S. The result is a product which is enriched in the heavier ³³S. Since the reaction is mass-dependent, the product will be even more enriched in the even heavier ³⁴S isotope. In fact, the product should be about twice as enriched in ³⁴S, since ³⁴S has two more neutrons than ³²S, while ³³S has only one extra neutron. Experiments show that this is approximately correct. δ³³S and δ³⁴S are related in a very simple, linear way for all mass-dependent isotope fractionations. If we plot δ³³S against δ³⁴S for a series of samples, we get a straight line with a characteristic slope (~1.94), no matter what combination of mass-dependent mechanisms caused the fractionation.

Variation from the characteristic mass-dependent line is expressed as Δ³⁴S. Any non-zero Δ³⁴S signals mass-independent sulfur isotope fractionation. If the plot of δ³³S against δ³⁴S does not yield a straight line, or if the slope is different from 1.94, then Δ³⁴S ≠ 0, and we have strong evidence of mass-independent fractionation -- and thus evidence of an oxygen-poor, sulfur-rich atmosphere, with deep UV penetration. (For a more carefully structured explanation of all this, see Baker, 2006).

MIF ends about 2300 Mya, thus free oxygen and an ozone layer must have been present. The point of all this is that mass-independent fractionation is detectable throughout the Archean Eon and in the Siderian period of the Paleoproterozoic. Farquhar et al. (2007). After a relatively brief transition (roughly, the Rhyacian) mass-independent sulfur isotope sorting disappears. Bekker et al. (2004). According to the sulfur data, the transition must have begun shortly after 2450 Mya and before 2330 Mya. Certainly by 2000 Mya (Orosirian Period), the atmosphere contained significant free oxygen. This does not mean lots of oxygen. It may have been as little as 0.0002% of the atmosphere (Holland, 2003) -- but enough so that near-surface UV-driven photochemical reactions were blocked by ozone and swamped by seawater sulfate reactions. ATW090224.

Iron Geochemistry

Oxidized iron is also found by 2300 Mya. This timing is generally confirmed from analysis of iron minerals. The 2320 My Timeball Hill Formation in South Africa, one of the units studied by Bekker's Group, contains "a large body of shallow-water hematitic ironstone ore," implying the presence of enough oxygen to oxidize iron to ferric oxide (Fe₂O₃). Holland (2003).

Again, this does not apply to the deep ocean. Holland points out that this oxidation may only apply to the ocean surface waters, as mentioned above. Unoxidized, banded iron formations, continued to form in deeper waters until about 1700 or 1800 Mya (later Statherian Period). Holland (2003); Poulton et al. (2004).

The later history of iron reflects slow oxygenation, not a sudden GOE. Rouxel et al. (2005) looked at iron isotope fractionation and believe that they can identify three "Iron Ages": (1) an Archean regime dominated by banded iron deposition, with the highly variable, generally negative, δ⁵⁶Fe characteristic of geothermal releases; (2) an intermediate stage; and (3) a post-1800 Mya period with higher, stable values characteristic of iron supplied by terrestrial weathering. However, as these authors note, there is an odd break in the deposition of banded iron between 2300 and 2100 (i.e., over the Rhyacian). After this, banded iron is formed again until the Statherian, but the δ⁵⁶Fe signal is now positive and less variable. Rouxel & Co. interpret this as signifying that the iron in Late Orosirian banded iron was dominated by terrestrial sources for some reason. Yet the isotopes remained incompletely homogenized. This seems odd, but it dovetails ever so sweetly with the results of Kump & Barley (2007) discussed earlier. This should be, partially, iron contributed from the spreading terrestrial LIPs --now exposed to a tiny, but increasing, amount of oxygen weathering. This is another important indicator that the GOE was somewhat slower and later than is usually supposed. ATW090224.

Carbon Geochemistry

Sulfur records the passing of an oxygen threshold -- somewhere in the vicinity of 10^-5 times the present atmospheric level. The carbon record suggests that this level continued to rise.

The GOE as measured by carbon isotopes. Carbon has two stable isotopes, ¹²C and ¹³C [5]. Enrichment in ¹³C can be expressed as δ¹³C, in the same way that an excess of ³³S can be expressed as δ³³S. There is some continuing debate about the fine points, but universal agreement that δ¹³C from marine carbonates shows an enormous peak (a "positive excursion") between 2300 and 2000 My (Late Rhyacian and Early Orosirian Periods). Possibly, this represents two or more closely spaced peaks. One common interpretation of this event is that it represents the creation of a great deal of free oxygen by photosynthetic bacteria. The assumption here is that such bacteria preferentially fixed light (¹²C) carbon into organic products (as they do today), leaving the atmosphere enriched in heavy (¹³C) carbon, which was then incorporated into marine carbonates. Holland (2003). Interestingly, the loss of mass-independent sulfur isotope fractionation precedes the positive δ¹³C excursion by at least 150 My. Bekker et al. (2005).

Carbon isotope fractionation doesn't necessarily imply widespread photosynthesis. Unfortunately, many natural processes can cause changes in δ¹³C. These processes include changes in: terrestrial weathering rates, ocean pH, temperature, tectonic activity, volcanic activity, availability of reduced metals, ocean depth, and ocean stratification, in addition to biological carbon fractionation. Even if we restrict ourselves to biological fractionation, we're in trouble. We have little idea of the biochemistry of organisms living more than two billion years ago. We associate biological carbon fractionation today with photosynthesis. Yet we have no guarantee that photosynthesis today follows quite the same biochemical pathways it used two billion years ago. As recently as the Miocene, a minor change in photosynthetic pathways in one plant clade (grasses) caused measurable shifts in δ¹³C [6].

We don't even have a guarantee that the dominant biological fractionation reaction was photosynthesis. It may have been, for example, fixation of methane and/or formaldehyde -- both of which may have been much more common at the end of the Archean than they are today. Goldblatt et al. (2006); Fuerst (2005). Recall also that any biogenic δ¹³C signal is a function of both biological activity and the selectivity of the biochemical reactions involved. In other words, the critters causing ¹²C depletion need not have been all that common, if their biochemistry was very efficient at separating light carbon atoms from heavy carbon atoms. Since the current data on Paleoproterozoic biochemistry are rather limited -- to say the least -- we simply don't know enough to draw confident conclusions.

Methane metabolism may be a better explanation for the carbon peak. One thing we do know is that methane metabolism has a much stronger tendency to fractionate carbon isotopes than does photosynthesis. Methane is also a stronger "greenhouse gas" than is carbon dioxide. For that reason, a good many researchers have played with the potential significance of methane for the GOE. The usual snowball crew have argued that the Late Archean was a "methane hothouse," and that the GOE supplied oxygen for a huge drawdown of methane, resulting in a global ice age, etc. Kopp et al. (2005). While we tend to be a little suspicious of global ice ages, we have no reason to doubt the methane drawdown. Thus the huge δ¹³C excursion of 2300-2100 Mya doesn't necessarily require an increase in oxygen much above the ~0.001% of present levels necessary to start forming an ozone shield. We suspect that global photosynthetic productivity was much lower than it is today. Perhaps the GOE was a very slow and dignified affair. The steps which appear geochemically sudden may only reflect the smooth and stately crossing of successive oxygen thresholds -- not sudden increases in oxygen.

A mathematical model suggests an initial slow increase in oxygen and a GOE. Then again, maybe not. Goldblatt et al. (2006) have devised an interesting mathematical model. Or, at least, it would probably be interesting if we had the vaguest comprehension of the mathematical underpinnings; but we don't. They look a bit like strongly non-linear differential equations fitted to some very rough empirical numbers -- but we're only guessing. Their bottom line is that predicted atmospheric oxygen has two stable solutions. Their assertion is that oxygen levels increased slowly and smoothly until the early Paleoproterozoic, when the ozone layer formed. This triggered a switch to a higher-oxygen regime (1-10% of PAL) with a transient increase or decrease in methane. One of the main problems with this model is that it cannot explain the geochemical iron record; but it does demonstrate the kinds of erratic atmospheric behavior which might have occurred as a result of initially slow increases in oxygen. ATW090224.

Calendar

The four periods of the Paleoproterozoic might be thought of in the following way.

Siderian: Fully modern plate tectonics; formation of ozone layer.

Rhyacian: GOE; hiatus in banded iron; spread of cyanobacteria with at least some (1 x 10^-5 PAL) atmospheric oxygen.

Orosirian: continental weathering develops; banded iron deposition resumes

Statherian: Sulfidic deep ocean; no banded iron deposition; eukaryotes (acritarchs) present.

Paleoproterozoic Climate

Possible "snowball Earth" episodes at 2400 and/or 2200 Mya. It has become customary to punctuate every discussion of the Proterozoic with exclamatory references to "Snowball Earth" glacial events, like a kind climatological Tourette syndrome. The Paleoproterozoic is no different. The current estimate among snowball aficionados is that the Earth froze up 3 times at various points in the Siderian ("Huronian" glaciations) and once in the Early Rhyacian, between 2200 and 2300 Mya ("Makganyene" glaciation). This is said to be coincident with, and probably related to, the GOE. Kopp et al. (2005). There seems to be some uncertainty about these dates. Kirschvink et al. (2000) originally dated the Makganyene glaciations to 2400 Mya. This corresponds to the Huronian glaciations of Kopp et al. (2005), but not to Kopp's Makganyene, which he dates to about 2200 Mya. A recent joint paper (Kirschvink & Kopp, 2008) does nothing to resolve the dating problem.

We are dubious about this. We are probably being excessively mistrustful and cynical [7]. However, we recommend a close reading of Eyles & Januszczak (2004) for a cold appraisal of snowball earth scenarios. Eyles & Januszczak's critique is largely delivered from a sedimentological perspective; and, if nothing else, you will learn a good deal of useful sedimentology. Their strongest point is that diamictites (poorly sorted conglomerates) and "lonestones" (anomalous large rocks surrounded by sediment layers) are not unique to glacial till and dropstones. See image. Their bottom line is that most of the evidence for snowball episodes is better interpreted as a heterogeneous collection of regional mass wasting events (landslides, earthquakes, and the like, including -- but not limited to -- glaciers). Eyles & Januszczak suggest that these may be closely connected with continental rifting.

Fortunately, most of the excellent work reported by both the Kirschvink and Kopp groups has much more relevance to the GOE and the evolution of cyanobacteria than to refrigeration or sedimentology. Accordingly, we will deal with it when we get to those issues. ATW090224.

Paleoproterozoic Life

Bacteria

Did photosynthetic cyanobacteria evolve before the Paleoproterozoic? Just about everyone agrees that the Paleoproterozoic was full of bacteria. Specifically, the Cyanobacteria ("blue-green algae") flourished and caused the Rhyacian (roughly) GOE. However, it is less clear whether cyanobacteria were products of the Paleoproterozoic or just happened to find the era unusually congenial. Evidence from biomarkers and "molecular clock" studies suggest that photosynthetic bacteria were present well back in Archean time.

The reader may be aware that we have a deep and unshakable suspicion of molecular clocks [8] and some shallower and more easily shaken doubts about "molecular fossils." If so, she will be unsurprised that we do not share this enthusiasm for Archean autotrophs.

Early evolution of photosynthesis is contradicted by the manganese record. On this point, the geochemical evidence collected by Kopp et al. (2005) is worth consideration. These authors point out that the world's earliest and largest manganese deposit cannot be older than 2220 My. This manganese is found as MnO₂, i.e. with manganese in the insoluble Mn⁴⁺ (Mn(IV)) oxidation state. Manganese in seawater at more-or-less neutral pH is in the highly soluble Mn²⁺ (Mn(II)) oxidation state. Oxidation from Mn II to Mn IV involves a hefty change in redox potential. That is precisely why manganese is used as the redox "battery" which drives photosystem II and produces free oxygen from water in cyanobacterial photosynthesis. In fact, according to Kopp et al. (2005), photosystem II is the only naturally-occurring reaction which could explain the geologically sudden accumulation of large amounts of Mn(IV) in one place.

Photosynthesis developed late from the coupling of two pre-existing cycles. This logic compels Kopp et al. to argue that photosynthetic cyanobacteria could not have evolved much earlier than the beginning of the Rhyacian (2300 Mya). This creates a problem, since the sulfur chemistry suggests an earlier date, as discussed above. Fortunately, all is well. One of the most plausible and detailed theories for the evolution of photosynthesis (Allen, 2005) would predict just such a chain of events. The essence is that photosystems I and II were originally independent, homologous anaerobic reaction systems. In fact, I- and II-like systems are found alone in various bacterial taxa. Alone, photosystem I-like cycles drive drive carbon fixation and oxidize sulfur. Photosystem II-like systems pump protons for ATP synthesis. Oxygen is produced, but it is largely recycled -- with some "leakage." Photosynthesis, as we know it in plants and cyanobacteria, is the reaction system which results when the two photosystems are coupled. When this happens -- abruptly (because it doesn't require much change) -- all of the redox potential from photosystem II becomes available to power carbon fixation. The oxygen from photosystem I is no longer applied to the cytochrome ATP-generating machinery and most can be discarded. Likewise, carbon fixation no longer requires sulfide.

Thus the chain of events involves the evolution and gradual spread of a bacterium with both photosystems in the Siderian or latest Archean. Note that this hypothetical bug is essentially an anaerobic beast, and the two photosystems are not coupled. As this bug spreads, it leaks trace amounts of oxygen (PSII) and oxidizes much of the available H₂S sulfur. However, by the Rhyacian, our bug starts to run out of hydrogen sulfide in the atmosphere and in surface waters. The redox state of the atmosphere has tipped toward oxidation. Continental weathering and outgassing from continental flood basalts (the source of terrestrial large igneous provinces) are pouring sulfates into the air and upper ocean. There is still only a trace of free oxygen, but enough to form an ozone layer. Thus, then, and perhaps only then, would it become strongly advantageous to take high-energy electrons directly from PSII and use them as input for PSI, discarding the oxygen by-product instead of recycling it. Hence: a long period of declining mass-independent sulfur isotope fractionation with only a trace of free oxygen, then a gradual increase in free oxygen and the appearance of oxidized manganese deposits [9].

The biomarker evidence for Archean bacteria was found to be contamination. What about those Archean biomarkers? Several papers by Jochen Brocks and others are frequently cited as supporting the presence of photosynthetic bacteria and eukaryotes. Brocks et al. (2003). This work was based on the study of hydrocarbons solvent-extracted from kerogen in shales of the Archean Pilbarra Craton in Australia. However, more recent work by a group of Australian geologists, again including Brocks, casts considerable, probably fatal, doubt on the original results. Rasmussen et al. (2008). The earlier papers had found complex organics which appeared to be breakdown products of chlorophylls and perhaps steroids. Rasmussen's spoilsports compared δ¹³C values of the extractable materials to the δ¹³C of the bulk kerogen. They also looked at microstructures within the shale. Their results showed that (a) biomarker δ¹³C was quite different from the δ¹³C of bulk kerogen organics in the rock; that (b) the relatively high (less negative) δ¹³C of the biomarkers was typical of organic compounds formed in the Paleoproterozoic or later; and (c) that the kerogen had probably formed under conditions which would have destroyed any biomarkers. Accordingly, the biomarker organics must have entered the shale from elsewhere -- in the Proterozoic, long after the shale was deposited. ATW090224.

Eukaryotes

The evidence for eukaryotes before the very late Paleoproterozoic is not strong. The evidence for eukaryotes in the Paleoproterozoic also looks moderately solid. Then again, so did hedge funds, before 2008. Like hedge funds, eukaryotes are a legitimate part of a balanced and diversified portfolio of ideas about the Paleoproterozoic. However, buying into this evidence -- or into hedge funds -- assumes that one is willing to accept ... certain risks. Notably, their value, in either case, depends strongly on what positions one has already taken and what everyone else in the market happens to believe about them during some given week.

Bengtson (2002: 293-294), wrote that "[t]he earliest fossil now commonly attributed to eukaryotes is the 1.85 billion-year-old Paleoproterozoic Grypania, a coiled, cylindrical organism that may attain half a meter in length and 2 mm in diameter. Because of its complexity and size, Grypania is commonly interpreted to be a eukaryotic alga." This seems to be the current consensus position. Knoll et al. (2006). Yet, as the image indicates, many workers attribute much older fossils to this genus. Han & Runnegar (1992). Some excellent images of these Michigan specimens may be found at James St. John's (Ohio State Univ., Newark) web site. To us, they look like fission tracks from a Proterozoic cloud chamber, or perhaps Burmese graffiti. Grypania is interpreted as a eukaryote solely because of its size and relative complexity. Bengston (2002). A few specimens have annular rings suggestive of cellular structures. Knoll et al. (2006). Those may or may not be good enough reasons.

Simple acritarchs were present by the end of the Paleoproterozoic. More convincing evidence comes only at the very end of the Paleoproterozoic (c. 1700 Mya), in the form of very simple plesiomorphic acritarchs in very low diversity (5-10 species). Knoll (1994); Huntley et al. (2006). These include Valeria lophostriata from China perhaps at ~1800 Mya and Australia 1650 Mya. Knoll et al. (2006). See image above from Hofmann (1999). These are round, organic-walled spheres, without processes. Bengston (2002). Acritarchs with distinct processes (Tappania) do not appear until slightly after 1500 Mya. Javaux et al. (2001).

The more credible "molecular clock" models also suggest a late -- perhaps very late -- Paleoproterozoic origin for eukaryotes. Yoon et al. (2004). Then again, we probably think this one is credible because it agrees with the other evidence. Even these very late Paleoproterozoic dates may be a shade too aggressive. Porter (2004). However, fossils of this age are so rare that we may never be any more certain. For now, a Statherian origin for eukaryotes seems the most reasonable guess. ATW090224.