91av

Plankton are some of the smallest living things on Earth, but they could
influence the climate of the whole planet. They are the most abundant form
of life in the oceans, in both weight and sheer numbers. A litre of seawater
can contain many millions of bacteria, thousands of phytoplankton-microscopic
plants-and hundreds of zooplankton-microscopic animals.

These tiny organisms influence our climate in several ways. They absorb
and scatter light, warming the topmost layers of the ocean, and they produce
volatile organic compounds, such as dimethyl sulphide, which help clouds
form. But their most significant role is moving carbon around the oceans,
on a scale large enough to affect levels of carbon dioxide in the atmosphere.
This is how plankton play a part in the natural greenhouse effect. No one
is yet sure how they will respond to the warming prompted by extra greenhouse
gases; they may help keep carbon dioxide within tolerable limits, or they
could thwart our best efforts to control its emissions.

The concentration of carbon dioxide in the atmosphere is increasing
at present, primarily because we are using fossil fuel, such as coal, oil
and gas. This is likely to lead to global warming and related changes in
the world’s climate. But we could expect far worse changes if all the carbon
dioxide produced by human activity stayed in the atmosphere: roughly half
of the annual emissions disappear, some 2.5 gigatonnes out of emissions
of between 6 and 7.5 gigatonnes of carbon. Until recently, scientists have
assumed that the oceans provide the sink for most, if not all, of the ‘missing’
carbon dioxide. But they have not made accurate measurements of the amount
absorbed, and no one really understands the processes involved.

This uncertainty is a cause for concern. The oceans store about 50 times
more carbon than the air, and, each year, oceans and atmosphere exchange
around 15 times as much carbon dioxide as human activities produce. If we
want to predict how carbon dioxide will build up in future, and, in turn,
understand changes in the climate, we need to understand how the ocean carbon
cycle works. This is the aim of the Joint Global Ocean Flux Study (JGOFS),
a 10-year programme that forms the main marine project in the International
Geosphere Biosphere Programme, which investigates environmental change on
a world scale.

JGOFS oceanographers, biologists and chemists first took to the seas-the
North Atlantic-for field studies in 1989. Their aim was to find out more
about how and why plankton grow, what then happens to the carbon that they
fix in their tissue, and what effect these organisms have on the exchange
of carbon dioxide between air and sea. One target was to find out exactly
what happens in the spring, when vast numbers of phytoplankton ‘bloom’ across
the oceans over a few weeks. This rapid growth can act as a feedback mechanism;
it can either speed up or slow down other changes in the physics and chemistry
of the seas. The researchers were particularly interested in measuring the
partial pressure of carbon dioxide at the surface of the sea, and finding
how much it varies from place to place in spring and summer-vital knowledge
for understanding how the oceans absorb and release this gas. Data researchers
already had suggested that partial pressure values varied smoothly both
over the course of the year and from place to place in the great ocean basins.

At the same time as the field studies, Pieter Tans, of the University
of Colorado, Inez Fung, of NASA’s Goddard Space Science Laboratory in New
York, and Taro Takahashi of Columbia University in New York State, were
developing models. They compared measurements of partial pressures of carbon
dioxide across the world’s oceans with data from models of atmospheric circulation.
They concluded that the oceans were much less important in absorbing carbon
dioxide than researchers had assumed; they estimated that the oceans absorbed
only between 0.3 and 0.8 gigatonnes of carbon each year.

Tans, Fung and Takahashi inferred that the carbon must be absorbed somewhere
on land, through some as yet unidentified process. But the JGOFS research
has revealed another possibility, based on detailed surveys of the oceans
carried out in the spring and summer of 1989 and 1990. The team consisted
of researchers from 15 laboratories, working in close collaboration with
JGOFS groups from the US, Germany, the Netherlands and Canada. They found
that, in the northeast Atlantic, to the south of Iceland and to the west
of the British Isles, the sea takes up widely varying amounts of carbon
dioxide at different places in the ocean. And this variation in absorption
related directly to the distribution of plankton, which they found was very
patchy across this sector of the seas (Nature, 8 March 1991).

In one area, from 45 degrees to 50 degrees North, around 1000 kilometres
west of Lands End, the peak reduction in the partial pressure of carbon
dioxide at the surface of the seas relative to that in the atmosphere was
about 70 microatmospheres (u atm) in late May and early June. That compares
with Tans, Fung and Takashashi’s estimate of a mean annual air-sea difference
of 15 u atm for the southern sector of the North Atlantic, based on summer
values of 1 u atm (from July to October) and winter values of 29 u atm (from
January to April). Further north, where existing data suggested that the
summer difference was around 37 u atm, Andrew Watson and colleagues at Plymouth
Marine Laboratory found that the difference could be as much as 200 u atm
near sea ice around Greenland, compared to the atmospheric level of around
350 u atm. For the North Atlantic, the annual averages that Tans and his
colleagues used seem to have underestimated the power of the sea to absorb
carbon dioxide. If other oceans are as variable as the results from JGOFS
suggest, the carbon budget might balance without an extra terrestrial sink.

But as well as measuring carbon dioxide in surface waters, the JGOFS
teams delved deeper into the mechanisms that carry carbon dioxide within
the oceans, finding records in sea floor sediments of the interplay between
oceans, the composition of the atmosphere and changes in the climate. Much
of what researchers know about this complicated system comes from the Earth’s
cycles of ice ages. Air bubbles trapped in polar ice have shown that there
was much less carbon dioxide in the atmosphere at the height of the last
ice age, which ended between 10 000 and 15 000 years ago. The slide into
an ice age starts because the Earth’s orbit around the Sun varies; this
has only a small direct effect on temperature but triggers other changes.
In particular about 200 gigatonnes of carbon disappear from the atmosphere
as the planet cools, only to return over a few thousand years at the end
of the glacial period. In round terms, this is the same amount of carbon
that human activities have added to the atmosphere over the past two centuries.

Where did the carbon dioxide go during the ice ages? The colder, drier
glacial climate would not have helped fix carbon in plants and soil on land.
Although there was more land, because the sea level was lower and the continental
shelves were exposed, these gains would have been offset by greater ice
cover at high latitudes and less rainfall at mid and low latitudes. The
carbon apparently went into the deep oceans. Data from marine sediments
have shown that the ‘biological pump’, acting via photosynthesis in surface
waters, became stronger and the deep water circulation more sluggish. More
carbon settled out of surface waters as organic particles, and its return
to the surface was slower.

Switching on an ice age

The researchers are especially interested in changes in ocean processes
near the poles, particularly the North Atlantic and Southern Ocean. These
regions probably acted as the switch for the change between glacial and
interglacial conditions. Today, there are far less phytoplankton growing
in the Southern Ocean than there could be; they do not use up all the available
nutrients such as nitrate and phosphate. Some scientists have suggested
that we could exploit this potential, and cut down on carbon dioxide in
the air, by artificially supplying trace elements such as iron, whose absence
might limit plankton growth .

If more phytoplankton grow, they can create a positive feedback mechanism;
during ice age cycles, cooler weather meant more plankton, and less carbon
dioxide in the atmosphere, which in turn led to further cooling. At the
end of an ice age, similar ocean and atmosphere feedbacks (working in the
opposite way) helped to warm the world. In addition, changes in ocean circulation
at this time carried more heat to high latitudes, acting either as the initial
switch, or as a very early response to orbital wobbles. This helped to release
more carbon dioxide from deep water, because the ocean then mixes more actively.

Apparently, the main effect of changes in marine productivity in the
past has been to destabilise the climate, reinforcing temperature trends
that were already under way. But the mechanisms and their implications for
the future are still uncertain. Are the biological controls on the movement
of carbon dioxide between air and sea more or less important than physical
factors? Are the oceanic responses that contributed to the long-term changes
in carbon distribution of the ice ages relevant to the current situation
of rapidly increasing carbon dioxide in the atmosphere?

These are questions that JGOFS is trying to answer. Although we need
many more detailed investigations of different ocean regions, the aim is
to partition reliably the net uptake of carbon between biological and physical
processes, and between marine and terrestrial sinks, without the need for
painstaking surveys over all ocean basins. The JGOFS measurements from the
North Atlantic agree with preliminary models developed by Arnold Taylor
and his colleagues at Plymouth Marine Laboratory that link the partial pressure
of carbon dioxide at a particular locality with temperature, depth of mixing
and abundance of phytoplankton in the upper ocean. Because satellites can
monitor temperature and chlorophyll-a plankton pigment-over large areas,
it should be possible in the future to estimate carbon fluxes across the
world by combining such data with information on sea conditions, cloud cover
and seasonal factors that relate to latitude.

Oceanographers are now designing large-scale field studies in collaboration
with the World Ocean Circulation Experiment to obtain the ‘sea truth’ data
needed to calibrate such a global monitoring exercise. But monitoring studies
require long-term commitment: for signals that are naturally noisy, it may
take twenty years or more before we can demonstrate statistically significant
trends and identify linked changes .

Even then, it might be difficult to separate cause and effect. If researchers
identify a trend in the past, this is not necessarily a reliable guide to
the future behaviour of such complex, nonlinear systems. If we want to make
predictions, we will have to understand the component processes in depth.

For that reason, the JGOFS programme includes investigations of what
happens to carbon throughout the water column, so that models can take full
account of these interacting subsystems. Without such information, any predictions
we can make today are speculative and have to be simplistic. For example,
we could assume that global warming will result in processes that now happen
at mid-latitudes moving towards the poles. On that basis, the expected changes
are hardly reassuring; a shift from cold to warm water regimes seems to
favour positive feedback mechanisms, which release more carbon dioxide.

Warmer waters, in general, are very stable near the surface, contain
phytoplankton of small cell size and are very efficient at recycling nutrients
and biogenic material in the upper ocean. As a result, relatively little
carbon sinks to deeper water.

In contrast, colder oceans mix much more in winter, renewing the supply
of nutrients to the surface layers. Phytoplankton productivity is greatest
in the spring, as light increases and the surface water warms and stabilises.
After this spring bloom, a pulse of plankton debris falls through the water
column, an effect enhanced by the larger cell size of phytoplankton in temperate
and near-polar oceans. Processes that help the debris clump together speed
up the rate at which it sinks.

Wolfgang Barkmann, of Kiel University, and John Woods, director of marine
sciences for the Natural Environment Research Council, developed a model
for these processes. It successfully simulates the main seasonal changes
in physical properties, their effect on the numbers of phytoplankton and
where they were in the water column.

The JGOFS North Atlantic programme looked closely at blooms and the
fate of carbon at sites of contrasting productivity along a north-south
transect in both 1989 and 1990. The researchers trapped material falling
through the water at particular depths, and took cores from the sea floor,
to determine sedimentation rates and the relationship of the sediment record
to long-term changes in ocean productivity and climate. In the upper ocean,
the aim was to balance the carbon budget for the sites and periods of intensive
study, with independent measurements of over 20 rates and processes.

The JGOFS teams are still analysing these data. But they have already
proved both significant and novel. In particular, the spring bloom itself
was much more complicated than expected. Instead of major changes in a few
species over large areas of the sea, there was a complex succession of separate
blooms in different communities of plankton. Furthermore, adjacent water
masses with subtly different physical and chemical properties bloomed in
distinct ways, at different times. Material in water near the ocean floor
showed that blooms in late summer also made a significant contribution to
the carbon flux to deeper water, but their scale and timing varied markedly
from 1989 to 1990.

The study also highlighted problems in quantifying the role of dissolved
organic carbon (This Week, 15 December 1990) and the need to take account
of the vertical migration of zooplankton. Investigations led by Martin Angel,
of the Institute of Oceanographic Sciences Deacon Laboratory, showed that
the abundant amphipod Themisto, has a big part to play in the rate at which
the biological pump works. It feeds at night on aggregates of phytoplankton
between 30 and 50 metres below the surface; at dawn, it dived between 200
and 300 metres, where it defecated. The effect of this energetic daily movement
is to speed up the flow of carbon from near the surface to the deep ocean.

Further surprises came from experiments and measurements in which the
research ship followed the flow of surface water, using drifting buoys tracked
by satellite. In 1989, US and German research vessels carried out several
studies of this type, each lasting between two and three weeks. In 1990,
the British JGOFS team was involved in a more ambitious experiment, using
the NERC research ships RRS Discovery and RRS Charles Darwin to follow a
marker buoy for nearly seven weeks from the end of April to mid-June.

Ten other buoys tracked the movement of a larger area of water, 10 000
square kilometres in all, where the researchers had surveyed the abundance
of phytoplankton and carbon dioxide parameters at the start of their study.
They released the marker buoy at a relatively quiet site at the edge of
a small eddy. But by the end of the study it had drifted more than 500 kilometres,
and the other buoys had spread over more than 260 000 square kilometres.
It was the equivalent of starting a study of a woodland in Norfolk and finding
one tree in Scotland within two months and the others scattered throughout
the rest of Britain. Furthermore, some of the leaves would be halfway to
the stratosphere, because the products of marine photosynthesis also move
vertically, at rates of around 100 metres a day.

Such dynamic complexity makes it essential that future JGOFS projects
pay as much attention to what goes where as to what grows where. Understanding
the rules that govern the biological and physical behaviour of carbon throughout
the world’s oceans is a scientific problem on a daunting scale. But we need
to know what turns the wheels, and, more importantly, what changes gear,
for the ocean carbon cycle: otherwise, major policy decisions on greenhouse
gas emissions will be made in ignorance of whether a temperature rise of
a few degrees might trigger, via plankton, a self-accelerating rise in carbon
dioxide. Or, if you prefer to be optimistic, whether global warming may
never happen.

Phillip Williamson and Arnold Taylor work on the NERC Biogeochemical
Ocean Flux Study (UK JGOFS) at Plymouth Marine Laboratory. John Gribbin
is physics consultant to 91av.

* * *

1: A technological fix that does not work

Debate is raging among marine biologists and climatologists over the
possibility that iron could provide the answer to the increased greenhouse
effect that human activities may have set in train. The prospect of such
a ‘technological fix’ stems from studies carried out by John Martin and
Steve Fitzwater, of the Moss Landing Marine Laboratory, in California.

They suggested that the growth of plankton at high latitudes is limited
by the amount of iron available in the sea. Carbon dioxide might be drawn
out of the atmosphere more effectively by adding some form of soluble iron
to the oceans in these regions. But the latest evidence indicates that even
if their suggestion is correct, there is no significant prospect of exploiting
this to reduce the build-up of carbon dioxide.

The notion that the growth of phytoplankton in Antarctic waters, at
least, might depend on the amount of iron available goes back half a century.
It was suggested by Sir Alister Hardy, perhaps best known today for his
idea that our ancestors spent a large part of their time in water-the ‘aquatic
ape’ hypothesis.

In the mid-1980s, Martin and Fitzwater tested the idea by adding dissolved
iron compounds to water from the northeast Pacific Ocean. They found that
this stimulated a rapid growth of plankton, until all the available nitrogen
in the water is used up (Nature, vol 331 p 341). Without the extra iron,
at least in test tubes, plankton stop growing when nutrients such as phosphates
and nitrates are still available-even if there is plenty of sunlight for
photosynthesis. The rationale behind this is that iron is an important ingredient
in many life processes, including the production of chlorophyll; without
chlorophyll the plants cannot photosynthesise.

Climatologists were intrigued. Air bubbles trapped in the Antarctic
ice show that during the most recent ice age there was less carbon dioxide
and more wind-blown dust in the air-dust that included iron. They reasoned
that dry, dusty winds blowing out of continental interiors, typical of an
ice age climate, might enrich the oceans with iron compounds. This would
encourage photosynthesis and draw carbon dioxide out of the atmosphere,
reducing the greenhouse effect in a self-sustaining feedback that cooled
the planet.

In February 1988, I suggested (Nature, vol 331 p 570) that if these
ideas were correct, we might be able to fix the present and future carbon
dioxide problem, by adding iron compounds to the ocean. Martin had come
to similar conclusions and proposed that a supertanker of iron could start
the next ice age.

That suggestion, made as he says, ‘more or less facetiously at a Journal
Club Lecture at Woods Hole Oceanographic Institution in July 1988’ (US JGOFS
Newsletter, April 1990, p 5) ran into a storm of opposition. Environmentalists
protested that such action would be ecological vandalism, likely to do more
harm (by disrupting natural ecological cycles) than good (by reducing carbon
dioxide).

Charles Miller, of Oregon State University, pointed out that if such
treatment was applied to the North Pacific, nutrients near the surface of
the ocean could be used up and biological activity would stop. This would
have catastrophic implications for fisheries, as well as for the uptake
of carbon dioxide. Things would only change when a large vertical mixing
event dredged up more nutrients from the depths.

Other researchers, including Karl Banse, of the University of Washington,
Seattle, argued that iron was not really a limiting factor on ocean productivity;
the effect applied only on the test tube scale-a conclusion hotly refuted
by Martin and Fitzwater. But just as this stimulating scientific debate
was getting into its stride, it seems to have run into a brick wall.

Tsung-Heng Peng, of Oak Ridge National Laboratory, in Tennessee, and
Wallace Broecker, of Columbia University, in New York, have modelled the
effects of adding iron to the seas. They conclude that ‘even if iron fertilisation
worked perfectly it would not significantly reduce the atmospheric carbon
dioxide content.’ From their model, after 100 years of totally successful
fertilisation, the amount of carbon dioxide in the atmosphere would be reduced
by only about 50 parts per million (Nature, vol 349 p 227).

At present, the carbon dioxide concentration is about 350 ppm. Even
if we make strenuous efforts to curb the growth in emissions of carbon dioxide,
atmospheric levels look set to increase to more than 500 ppm over the next
100 years. Even worse, over the same period the enhancement of the greenhouse
effect by other gases such as chlorofluorocarbons and methane looks set
to be more than the increase due to carbon dioxide alone.

Iron fertilisation may be a key to understanding natural climatic changes
of the past, but does not seem to be the answer to the problems posed by
the anthropogenic greenhouse effect. The effort would be better devoted
to other aspects of the greenhouse problem.

2: Plankton and the Gulf Stream

If we want to find out more about plankton and the climate, we need
a year by year record, which the Continuous Plankton Recorder survey (CPR)
provides. Established in the 1930s by Sir Alister Hardy, the CPR has surveyed
the northeast Atlantic each month from 1948 to the present day. The survey
is now an independent Foundation with support from the Ministry of Agriculture,
Fisheries and Food (MAFF) and an international funding consortium.

The CPR records show that the numbers of phytoplankton and zooplankton
in the northeast Atlantic declined from 1948 to 1980; since then, the trend
has reversed. Michael Colebrook, of Plymouth Marine Laboratory, and Bob
Dickson, of the MAFF at Lowestoft, have shown that these trends were the
result of natural environmental factors. In particular, wind strength was
more important than temperature.

Ten years ago we identified another environmental factor that correlated
with data from the CPR-the position of the Gulf Stream off Cape Hatteras,
between 65 degrees and 80 degrees West, after it leaves the US coastline.
In years when the northern boundary of the Gulf Stream lay farther north,
the CPR data showed more copepods (a major group of zooplankton) in sea
areas west and northwest of Britain.

But at that time there was data on the position of the Gulf Stream for
only 12 years; we thought of the link as a statistical curiosity. With the
help of Nick Baker, John Stephens and Michael Colebrook, I have now analysed
the data again with extra care and extended the time series to cover the
24 years from 1966 to 1989.

We found that the revised correlations between annual averages in the
north-south position of the Gulf Stream and the abundance of plankton were
statistically significant for many more groups of plankton than before,
and applied to the North Sea and the open Atlantic. Furthermore, the relationship
was remarkably good, considering the problems in taking samples. These arise
because the distribution of the plankton is patchy, and it is easier both
to collect and identify larger species. The position of the Gulf Stream
also varies considerably from month to month; this was smoothed out in the
annual averages.

While more detailed interpretation is now going on, we can draw several
important conclusions. First, the abundance of plankton in the North Atlantic
responds to subtle climatic and oceanographic factors that operate over
many thousands of kilometres. Secondly, and as a result of this, many of
the changes in biological activity from year to year are not merely the
result of chaotic dynamics within the upper ocean. Thirdly, it follows that
we should be able to determine the main climatic factors controlling the
ocean carbon cycle and predict its future behaviour.

The CPR data set varies considerably; for example, in the west of the
central North Sea, there was a six-fold decline in the abundance of copepods
from roughly 1950 to 1980. These results arose under a relatively stable
climate-we can expect global warming of a few degrees to have a much greater
effect, with the possibility of a synchronised response worldwide.

Arnold Taylor