Oxygen
Page 12
Such catastrophic scenarios are only too familiar to the environmental scientists who try to model changes in atmospheric composition, but there are difficulties with more subtle forms of negative feedback too.
One mechanism, proposed in the late 1970s by Andrew Watson, Lovelock and Margulis as part of the Gaia hypothesis, suggested that bacterial methane production might stabilize oxygen levels.
Methane-producing bacteria thrive in stagnant swamps where oxygen levels are very low, and they cannot tolerate higher levels. They gain their energy by breaking down organic remains in the swamps to
80 • THE BOLSOVER DRAGONFLY
release methane gas. This is no trivial process. Lovelock estimates that some 400 million tons of methane are emitted into the atmosphere each year by swamp bacteria (industrial pollution, farming and landfill sites have now more than doubled this figure, contributing to global warming, among other things). The theory goes that if the burial of organic matter in swamps were to increase, pushing up oxygen levels, new colonies of methane bacteria would thrive on the putrefying detritus, ultimately belching out excess methane into the surrounding air. Methane escaping from the swamps reacts with oxygen over a period of several years to form carbon dioxide, thereby lowering oxygen again. Conversely, a lower burial rate would support a smaller population of methane bacteria, which would emit less methane and so encourage oxygen levels to rise again.
The continuous feedback operating in this cycle might prevent large fluctuations in oxygen levels. The difficulty with the theory is that it predicts that methane bacteria should maintain permanent carbon burial at a roughly constant rate, as the bacteria regulate oxygen levels by breaking down organic matter that would otherwise be permanently buried. In fact, there are periods when the geological record shows that it varied substantially. The massive amount of coal formed in the Carboniferous and early Permian was clearly not broken down by methane bacteria in coal swamps, so we can only conclude that there were times when the methane cycle was insufficient to regulate atmospheric oxygen.2
More convincing as a biological feedback mechanism for regulating oxygen is a curious phenomenon that affects plants, suppressing their growth and productivity; in some circumstances, plant growth can be halted completely. The phenomenon is known as photorespiration and, unlike the plant’s normal mitochondrial respiration, takes place only in sunlight. Its purpose is a mystery. The net effect is that the plant takes up oxygen and releases carbon dioxide, which parallels normal respiration (hence the name) but fails to generate any energy. Also unlike normal respiration, photorespiration competes with photosynthesis for the use of an enzyme known by the sonorous acronym of Rubisco (which stands for ribulose-1,5-bisphosphate carboxylase/oxygenase). This contest for the 2 Availability of nutrients such as phosphate has also been proposed to limit the burial rate, but since the ratio of phosphorus to carbon is much lower in land plants than in marine algae and plankton, more carbon can be buried on land per unit of phosphorus. Phosphate abundance is therefore less likely to affect burial rates in terrestrial, compared with marine, environments.
Oxygen and the Rise of the Giants • 81
affections of Rubisco undermines the efficiency of photosynthesis and so reduces plant growth.
Rubisco is the enzyme that binds carbon dioxide and incorporates it into carbohydrate in photosynthesis. It is often justifiably claimed to be the most important enzyme in the world. Certainly, weight for weight, it is the most abundant enzyme on Earth. Without Rubisco, photosynthesis as we know it could not take place. With it, we have a different problem.
By enzyme standards, Rubisco is not at all discriminating. A molecular two-timer, it binds oxygen almost as eagerly as it does carbon dioxide.
When Rubisco binds carbon dioxide, its lawfully wedded wife, the plant uses the carbon constructively to make sugars, fats and proteins. If, however, Rubisco binds mistress oxygen, then a large number of enzymes start catalysing a useless chain of biochemical reactions, the result of which is a return to square one. This energy-sapping chain of reactions slows down plant growth as surely as a politician’s misdemeanours retard his rise to power.
The rate of photorespiration increases with both the temperature and the oxygen level. This means that plant growth grinds to a halt in hot, oxygen-rich conditions. Even in normal air, the apparently pointless squandering of resources can reduce growth by as much as 40 per cent in tropical areas. There is a corresponding drain on agricultural productivity, though this may be masked to the casual gaze by ameliorating factors such as rainfall, soil fertility and length of growing season.
Despite its apparent futility, photorespiration is almost universal among plants, although many have evolved ways of reducing its detrimental effects.3 It has been maintained by evolution for some reason. In other words, it must be useful for something, otherwise it would have been unceremoniously dumped in the struggle for survival. This premise is backed by the failure of numerous commercial attempts to breed plants that do not photorespire, often motivated by an honest desire to increase productivity in developing countries. Curiously, none of these genetically modified plants could survive in normal air, yet most could thrive in air rich in carbon dioxide but low in oxygen, suggesting that photorespiration might protect against oxygen toxicity in some way. This would 3 Photorespiration is a particular problem for so-called C3 plants, which include most trees and shrubs. Grasses are mostly C4 plants, and escape the worst excesses of photorespiration by compartmentalizing their photosynthetic machinery. They capture carbon dioxide and then release it in large amounts into the cellular compartment containing Rubisco. In these circumstances CO2 out-competes O2 for the services of Rubisco.
82 • THE BOLSOVER DRAGONFLY
explain why photorespiration is less necessary at low oxygen levels, and more necessary at normal or high oxygen levels. Whatever the reason, the upshot is that photorespiration stunts plant growth if atmospheric oxygen levels are high.
The potential magnitude of photorespiration fits it as a plausible mechanism for stabilizing atmospheric oxygen. If oxygen levels were to rise, the rate of photorespiration would rise in tandem and plant growth would falter. Stunted plants would produce less oxygen by photosynthesis, thereby adjusting atmospheric oxygen back to previous levels. An elegant feature of this hypothesis is that it does not imply that carbon burial should be constant. Quite the contrary. The rate of burial should in principle vary with the exuberance of plant growth: if there is no growth there can be no burial, and vice versa. The big question is empirical — can photorespiration really account for both oxygen regulation and variations in carbon burial?
The answer is by no means certain, but the question is at least open to experiment. Some studies suggest that, while surely playing a role, photorespiration alone cannot maintain atmospheric oxygen at a constant level all the time. In coming to this conclusion, in an important 1998 paper in the Philosophical Transactions of the Royal Society, David Beerling and his colleagues at the University of Sheffield measured the growth of a variety of plants at incremental oxygen levels from 21 to 35
per cent. At 25°C, overall productivity was down by 18 per cent in oxygen-rich air compared with normal air, confirming an effect on growth rates. However, the magnitude of this effect was not the same for all plants: evolutionarily ancient groups of plants fared much better than their modern cousins. Plants that had evolved during the Carboniferous, such as ferns, gingko and cycads (palm-like evergreens with cones rather than coconuts) were less sensitive to increased oxygen than were plants that had evolved more recently, such as the angiosperms — the largest group of plants today, which include deciduous trees and shrubs, our staple crops, and all other herbaceous crops and flowers. The more ancient groups of plants were also more likely to adapt to the new conditions by changing the structure of their leaves. In particular, these plants increased the number of stomata (the pores through which gases enter and leave the leaf), allowing more carbon dioxide to accumulate insid
e the leaf.
Interestingly, if carbon dioxide levels in the air were doubled, from 300 to 600 parts per million in these experiments, plant growth was not
Oxygen and the Rise of the Giants • 83
diminished at all, and indeed productivity sometimes rose. Although carbon dioxide levels generally fall as oxygen rises, most geologists agree that carbon dioxide fell from a high point of about 3000 parts per million in the Devonian (385 million years ago) to a low point of 300 parts per million by the end of the Permian (245 million years ago) (Figure 5). Carbon dioxide levels may therefore have been higher in the Carboniferous than today. All in all, the Sheffield team concluded that high oxygen levels during the Carboniferous and early Permian could have done little more than thin plant growth in tropical areas.
Swamp bacteria, nutrients and photorespiration may well help to regulate oxygen levels under normal conditions, but could at best only blunt the Oxygen
35
30
25
PAL
20
15
Per cent oxygen
10
5
E
O
S
D
C
P
Tr
J
K
T
Carbon dioxide
0.5
0.4
0.3
0.2
0.1
Per cent carbon dioxide
PAL
E
O
S
D
C
P
Tr
J
K
T
–600
–500
–400
–300
–200
–100
0
Time (million years ago)
Figure 5: Changes in atmospheric composition over the Phanerozoic period, from 600 million years ago, based on the models of Robert Berner. Oxygen levels (top graph) reached a peak of 35% in the late Carboniferous and early Permian, before falling to 15% in the late Permian. Oxygen levels peaked a second time at 25 to 30% in the late Cretaceous (K) before falling to present atmospheric levels in the Tertiary (T). Carbon dioxide levels (bottom graph) fell from 0.5% in the Silurian (S) to around 0.03% by the end of the Carboniferous. Reproduced with permission from Graham et al., and Nature.
84 • THE BOLSOVER DRAGONFLY
large changes in atmospheric oxygen predicted by the high rate of carbon burial in the late Carboniferous and the early Permian. Perhaps it is time to look a little more closely at the riddle posed by this 70-million-year period, which lasted from around 330 to 260 million years ago. Ninety per cent of the world’s coal reserves date to a period that accounts for less than 2 per cent of the Earth’s history. The rate of coal burial was therefore 600 times faster than the average for the rest of geological time. Most organic matter is not buried as coal, of course (see Chapter 2), but systematic analyses of the organic content of sedimentary rocks all over the world confirm that the total amount of organic matter buried during the Carboniferous and early Permian was much greater than during any other period, including the present day.4
Unique events are best explained by singular circumstances. The most believable explanation for the high rate of carbon burial during the Carboniferous and early Permian invokes an accidental alliance of geology, climate and biology. Two factors in particular were probably important. First, the continents had recently converged to form a lowlying supercontinent called Pangaea, and the wet climate and vast flood plains could hardly have provided a better nursery for coal swamps.
Second, the rise of large woody plants, the first trees, about 375 million years ago, brought about the verdant colonization of upland areas, as well as the swamps and seashores. Woody plants depend on lignin for structural support. Even today, bacteria have difficulty digesting lignin, but during the Carboniferous and early Permian there must have been a huge discrepancy between the amount of lignin formed by woody plants and the amount broken down by swamp bacteria.
The unparallelled rate of coal formation in the Carboniferous and early Permian is explained, then, by an exceptionally high rate of lignin production, an exceptionally low rate of lignin breakdown, and nearly perfect conditions for preserving organic matter on an unprecedented scale. We know of no negative feedback mechanism that could have constrained atmospheric oxygen under these conditions, so we can only conclude that oxygen levels must have gone up, and probably quite sub-4 In addition to carbon compounds, we must also consider the burial of iron pyrites, or fool’s gold. When hydrogen sulphide is formed by sulphate-reducing bacteria, it can either react with oxygen to form sulphates, or if no oxygen is present it may react with dissolved iron to form iron pyrites. If iron pyrites are formed and then buried, some oxygen that would otherwise have been consumed by hydrogen sulphide is instead retained in the atmosphere.
Simultaneously high rates of carbon and pyrite burial therefore translate into the highest rate of oxygen formation.
Oxygen and the Rise of the Giants • 85
stantially. Like M. G. Rutten, I’m quite satisfied by this line of evidence; but the question remains, how much is a lot?
The balance sheet for photosynthetic oxygen production shows that a fixed amount of oxygen is left in the air for each equivalent unit of organic carbon buried (see Chapter 2, page 24). In principle, to calculate oxygen levels all we need to know is how much organic matter was buried in the past. From this figure we must subtract the amount of buried matter that was later exposed by erosion and returned to the atmosphere as carbon dioxide. On our balance sheet, carbon returned to the air through erosion is no different from carbon burnt for energy and returned immediately to the air as carbon dioxide. The difference in rate, however, is critical. Coal that was buried in the Carboniferous, and is today dug out and burnt, was nonetheless buried for 300 million years. Its burial helped raise atmospheric oxygen levels throughout this time, just as burning it is lowering them again today (albeit by a matter of 2 parts per million per year, against a background level of 210 000 parts per million).
To put numbers on rates of carbon burial and erosion at times in the distant past might seem a reckless intellectual escapade, but with a little hedging the Yale University geochemist Robert Berner and his former doctoral student Donald Canfield have succeeded in assigning some reasonably sensible parameters. They argued that, because the great bulk of organic matter is buried in coal seams, silting river estuaries and shallow continental shelves, we do not need to worry too much about rocks that formed at the bottom of the deep oceans. It is then a straightforward, if soul-destroying, task to determine the relative abundance of the different sorts of continental sedimentary rocks, as a quick glance at any detailed geological map will testify. The organic content of these rocks can be measured directly. The real difficulty comes in calculating differential rates of erosion. We can assume that older rocks are more likely to have been completely lost by erosion or metamorphism, whereas younger rocks, buried closer to the surface, are more readily exposed and eroded.
Another factor to take into account is whether burial originally occurred in places with good potential for preservation, such as coal swamps (as occurred widely in the Carboniferous), compared with burial at sites subject to high rates of erosion, such as alluvial plains (which was more common in the Permian).
By estimating rates for carbon burial and rock erosion on the basis of the evidence available, Berner and Canfield calculated the apparent changes in oxygen levels over the past 600 million years. They came up
86 • THE BOLSOVER DRAGONFLY
with a graph that sent shock waves through the geological establishment.
Oxygen levels, they said, rose to 35 per cent during the late Carboniferous and early Permian, then fell to 15 per cent in the late Permian, to cause the worst mass extinction ever recorded. Later, during the Cr
etaceous (the final age of the dinosaurs), oxygen levels crept up again, this time to around 25 or 30 per cent (see Figure 5).
However impeccable the logic, numbers such as these strain credibility, and stand opposed to most people’s intuitive feeling. Perhaps for this reason, conclusions about ancient atmospheres based on computer modelling meet continuing resistance. Most scientists are distrustful of mathematical or philosophical reasoning unsupported by empirical observation — memorably dismissed as ‘fact-free science’ by the guru of evolutionary biologists, John Maynard Smith. A famous example of fatuous logic is the conundrum posed by an ancient Greek, Zeno of Elea, which troubled logicians for centuries but can rarely have cost a scientist a night’s sleep: movement is impossible, said Zeno, because to complete a single step one must first complete half a step, then half of the remainder, and so on in exponential fashion. Just as an exponential curve does not touch the baseline until infinity, so the infinite number of half steps precludes the possibility of ever taking a whole step. Berner’s and Canfield’s models may be far removed from the perverseness of Zeno’s paradox, but even though their calculations are based on empirical data, such an apparently improbable outcome will always invite the reply that an important factor has simply been overlooked.
The only really satisfying way of confirming or rejecting the hypothesis that oxygen levels once reached 35 per cent is to measure them: could there be, somewhere, a pocket of ancient air, miraculously undisturbed for hundreds of millions of years? The idea seems outrageous, but polar scientists have been drilling cores of ice from deep within the Arctic and Antarctic ice caps over many years, in an effort to read the preserved record of environmental change. The results have revealed much about the speed and magnitude of climatic changes in the past, as well as the extent of industrial pollution in Roman times and more recently.