ÐÓ°ÉÔ­´´

THE SPRING THAW is a momentous event in the cycle of many ecosystems.
When the liquid that flushes from the snow is acidic, the consequences for
the life in lakes and streams can be serious. Usually, the most important
source of the acid that finds its way into water courses is the soil; acids
collect there during the winter, when there is little movement of water
through the soil. In spring, water begins to percolate through the warming
soil and flushes out the accumulated acids. More insidious a source of acid
is the snow itself. As it lies on the ground, snow acts as a reservoir,
taking pollutants from the atmosphere and holding them in storage. Even
snow that looks clean and pure can contain a wide range of contaminants.
And, when the snow begins to melt, rivulets and streams of acidic water
flow over the ground and through the soil into water courses.

At midwinter snow covers around 40 per cent of the land of the northern
hemisphere – as much land as was permanently ice-covered during the most
recent ice age. This seasonal blanket lasts for some 120 days a year, and
reaches its deepest just before the spring melt. At this time it is, on
average, a little more than half a metre deep. There is considerable variation
from place to place. In parts of the Scottish Highlands, for example, ‘seasonal
snow cover’ can last all year and reach a depth of 15 metres. Where snow
is a regular feature of the year, ecosystems have evolved to cope with a
hydrological cycle that veers from winter drought, when water is locked
up in ice and snow, to cascades in spring when the snow melts.

Snow builds up at just the time of year when the concentrations of pollutants
in the atmosphere are at their highest. As droplets of water and ice crystals
form in clouds, they incorporate acid gases and aerosols – minuscule blobs
of acid and dust. As snowflakes develop and fall, they sweep out other polluted
droplets and ice crystals and collide with aerosols and larger particles
of pollutants, carrying them to the ground. The snowpack acts as a temporary
reservoir, locking up the pollutants for several months. When it thaws in
spring, the meltwaters remove the chemicals from the snow and deliver them
to the underlying soil, vegetation, water courses and lakes – all in the
space of a few days.

Recent research has highlighted the way that changes in the chemistry
of snow, and therefore the composition of the meltwater, can disturb susceptible
ecosystems. Many scientists believe that the spring pulse of pollutants
discharged from the melting snow, the ‘acid flush’, is growing worse because
of increasing atmospheric pollution, which now reaches relatively remote
snow-covered regions. The Arctic haze, for example, is formed by pollutants
carried from Eastern Europe and the Soviet Union.

ÐÓ°ÉÔ­´´s in Scandinavia, Western Europe and North America recognise
the importance of the springtime acid flush. There is good evidence linking
it to another event that is most dramatic in Scandinavia: the ‘fish kill’,
the occasional appearance of a mass of dead fish in a lake or river immediately
after the spring thaw. During the melt, the pH of water in streams and lakes
plummets. At the same time, the concentration of free aluminium increases.
Tests in the laboratory on fish show that high concentrations of free aluminium,
coupled with low pH and low concentrations of calcium, damage their gills.
Aluminium is normally present in high concentrations in the soil: it becomes
mobile when acidic waters flow through the soil and the beds of streams.

There are few direct observations of fish kills during the period of
spring melt in Britain or North America. But in view of these results and
the chemical changes in stream water when the snow melts, some researchers
believe that stocks of fish might be damaged in some parts of upland Britain
and North America, and there might even be mass mortality. Direct evidence
for this is hard to come by, however. This is because the period of flushing
is so short and because the young stages of fish, which are difficult to
observe and to sample, are likely to be affected most. Other, less obvious,
effects are still more difficult to measure. The sudden influx of acid to
the water could simply sap the vitality of fish, so that they succumb to
disease more easily.

There is another fundamental problem in assessing cause and effect during
episodes of fish kill, or indeed in trying to link the chemistry of the
snow with that of stream water. In many water catchments most of the extra
water discharged into streams and rivers during the thaw is ‘old’ water,
displaced from the lower levels of the soil and aquifers by the ‘new’ meltwater.
The discharge of old water through the water courses on the surface is often
associated with a lowering of pH and increased concentrations of free aluminium.
Thus, the chemical changes in stream water that might trigger a fish kill
are not necessarily directly related to the chemistry of the waters melting
from the snowpack. Nevertheless, on some occasions and in some places, the
meltwater does contribute a large amount to the discharge of a stream. This
happens when the route from snow to stream is short: when the water table
is high, for example, meltwater can enter a stream within hours. And when
the soil is saturated and frozen, meltwaters run straight over it and into
the nearest stream. Melting banks of snow along streams and the shores of
lakes also flush directly into the water.

Fish kills aside, polluted meltwaters can damage the ecosystem in several
ways. The soil underlying seasonal snow is in the front line: microbial
communities in the soil are especially vulnerable and can be badly disrupted
by acid flushes. Ian Thompson, of the University of East Anglia, found that
acid meltwater can reduce the number of bacteria in the upper layers of
some alpine soils to a thirtieth of their usual number. This may be partly
compensated by an ncrease in numbers, about tenfold, in the lower levels
of the soil, probably as a result of the transport of nutrients down through
the soil. The structure of the bacterial community also changes; for instance,
the number of Micrococcus plummets while Pseudomonas increases.

What goes in must come out

The acid flush illustrates how important the chemistry of melting snow
is to the stability of an ecosystem. Chemists and glaciologists have studied
polar snow intensively for decades. It never melts, and holds a record of
the atmospheric conditions in the past few millennia and gives some indications
of what is happening in the atmosphere today . The researchers also assume
that no chemical reactions of any significance take place within the snowpack.

Seasonal snow is another matter. Until recently, when the fish kills
aroused interest, almost nothing was known about what went on chemically
inside the snow. What little research had been done was restricted largely
to studies of what goes into the snow pack and what comes out when it melts.
Even these relatively simple studies are fraught with problems. It is extremely
difficult to catch a representative sample of falling snow. And in some
mountain areas recently fallen snow varies widely in its chemical composition.
Even in one place, the chemistry of individual snowfalls can vary greatly.
Sampling over four months on Cairngorm Mountain in Scotland, a group of
researchers from the universities of East Anglia, Southampton and Aberystwyth
collected freshly fallen snow each day. The pH varied hugely. The most acidic
was highly polluted black snow, with a pH of 3 – about as acid as vinegar.
Most of the carbon-rich particles that made the snow black originated in
the polluted atmosphere of Eastern Europe. At the other end of the scale,
an orange snow, coloured by North African sand, was an alkaline pH 8, on
a par with baking soda. In the Alps occasional deposits of alkaline dust
(which fall in red snow) may buffer the acid waters that flush from melting
snow.

Falling snow scavenges pollutants from the atmosphere, a process known
as ‘wet deposition’. But once the snow is lying on the ground it can accumulate
more pollutants, from gases, aerosols and particles in the atmosphere, in
a process known as ‘dry deposition’. The process is hard to measure because
the concentrations of pollutants in the air tend to be low, and the amount
actually deposited very variable. A cold covering of snow containing a small
amount of liquid water receives little sulphur dioxide compared with dry
deposits on vegetation. But, in the few days before the meltwater leaves
the snowpack, grains of snow begin to melt and the amount of water in the
snow increases. The snow is said to be ‘ripe’. At this stage, dry deposition
of highly soluble sulphur dioxide increases enormously.

Attempts to measure the total amount of chemical within the snowpack
are also beset with problems, especially in mountain regions. One is that
the wind redistributes the snow and mixes the layers. Moreover, chemical
changes take place within the snow even before the period of rapid change
just before the melt begins.

Once the thaw begins, the gradual chemical evolution of the snowpack
gives way to a rapid loss of solutes each time meltwater flushes out. Scandinavian
scientists were the first to observe that the first meltwaters to leave
the pack carry a higher concentration of solutes than the snow they leave.
Typically, between 50 and 80 per cent of the total amount of solutes in
the snow are released in the first 20 to 30 per cent of the meltwater. This
is called the concentration effect. The average concentration of soluble
impurities in this fraction of the meltwater is typically two to two-and-a-half
times the initial concentration in the snowpack, but the very first water
released can contain 10 times the initial concentration. Several factors
contribute to this. First, snowflakes begin to change shape almost immediately
after they fall. Smaller ice crystals coalesce and grow into larger crystals
over several months. The solutes originally distributed throughout the snowflakes
migrate to form thin films on the surface of each larger ice crystal or
grain of snow. The concentrated solutions accumulate at the junctions where
three grains meet in a ‘ripe’ snow. As the snow melts, these pockets of
concentrated solutions are the first to flow out of the snowpack in the
meltwater. When snow melts slowly, there is time for more of the solute
to diffuse from these junctions into the channels of meltwater, resulting
in relatively high concentrations of impurities.

A second reason why some meltwater is more concentrated is that a deep
blanket of snow may contain many layers from different falls of snow. Some
falls are more polluted than others; for example, a layer of black snow
may be sandwiched between fairly clean snow. The polluted bands contain
more solutes than the clean ones, and they contribute more solutes when
they melt. Similarly, snow may melt and refreeze, producing a band with
more concentrated impurities. If such a band is close to the base of the
snowpack, it can deliver especially high concentrations of solutes to the
soil beneath it when the pack eventually thaws.

To complicate matters further, the chemicals in the snow separate out
at different rates, entering the meltwater at different times. Experiments
in the field and in the laboratory have shown that such ‘preferential elution’
occurs with sulphate and nitrate ions, for instance. This effect is difficult
to study because the drainage channels that crisscross the snowpack may
deliver a mixture of meltwaters to the water course, masking any differences.
Although this variable rate of washing out of different solutes seems to
be a universal and important control on the chemistry of the snowpack during
the few days or weeks of melt, the phenomenon is rarely reflected in the
chemical composition of the streams that receive the meltwater. This may
be because the stream receives water from different zones of the snowpack,
which might be melting at different rates; because of interactions between
the water and the soil or the bed of the stream; or because of contributions
from ‘old’ water.

Where the meltwaters make a large and direct contribution to streams
and rivers, we need to understand what governs these concentration factors
during snowmelt. To this end, researchers in North America and Britain are
looking not just at the processes that are taking place at the time of the
thaw, but also at those that condition the chemical composition and structure
of the snowpack before it melts.

One of the first hints that these processes are important came when
chemists, including Gerry Jones in Quebec, looked at the fate of nitrogen
compounds, such as ammonia and nitrate, in snowpacks in northern forests.
They found that the snow had less nitrogen when it melted than when it had
first fallen. The decrease, they concluded, had to be the result of microorganisms
living off organic debris, even at relatively low temperatures. The main
burst of microbial activity takes place during the spring melt, however,
when nutrients carried down through the pack stimulate bacteria, algae and
fungi to grow. So, as the population of microbes grows, they consume the
nutrients within the snow – at the same time altering the pH, and the mobility
of trace metals within the snowpack.

Another important factor could be light. Sometimes the concentration
of ozone over Alpine snowpacks is very high, and some researchers believe
that this results from photochemical changes to the appreciable amounts
of hydrogen peroxide in the snow. The amount of hydrogen peroxide seems
to diminish as the snow ages, probably because light causes the hydrogen
peroxide to split in two. The oxygen atoms released combine with oxygen
molecules in the air to produce ozone.

Light could also excite other important chemicals in the snow, and so
trigger photochemical processes. In the presence of ultraviolet light, for
example, the nitrate ion induces the hydroxyl radical to form. These radicals
are highly reactive and rapidly combine with many organic compounds. Like
hydrogen peroxide, they are strong oxidising agents, that is, they will
remove electrons from other molecules.

Snow crystals as catalysts

There is now strong evidence that the levels of these substances and
their precursors, such as nitrate, fluctuate in snowpacks. This could be
due to exchanges between the snow and the atmosphere, or to processes going
on in the snowpack. With the help of light, the snowpack could be reducing
the damage to the environment by breaking down toxic micropollutants, such
as polyaromatic hydrocarbons, to smaller and less toxic molecules. These
would be trapped in the snow and decompose before the spring melt.

We do not yet know what role ice crystals play in these snow-bound chemical
reactions. In the atmosphere, ice crystals are important mediators of photochemical
reactions involving the conversion of chlorine nitrate and hydrogen chloride
into chlorine and nitric acid. Light then decomposes the chlorine molecules
into highly reactive chlorine atoms that trigger the breakdown of ozone
to oxygen.

The ice crystals seem to provide an active surface that concentrates
hydrogen chloride and chlorine nitrate. Both these chemicals are then ideally
set up to react, producing chlorine and nitric acid. Chemists would recognise
the ice crystal as a heterogeneous catalyst (a solid surface that promotes
the reaction of two gases) like the kind used in industrial reactions. The
ice crystal acts as a catalyst by accelerating the production of chlorine
and chlorine radicals. We can only speculate that similar processes occur
on ice crystals in snow on the ground.

The potential importance of these processes is reflected by the growing
collaboration between snow chemists of many nations. Studies of how the
melt affects the environment depend on an understanding of what happens
within the snow. Although we cannot yet assess the extent and precise cause
of the worst ecological upset we know about, the spring fish kills, the
potential scale of the problem is clear. In a recent survey of the lakes
of southern Norway, the country’s State Pollution Control Authority classified
60 per cent of the lakes as ‘acidified’. In spring, the number rises as
acid flushes pour into the lakes. Researchers have also reported an increase
in the number of acidic lakes in the Canadian Shield during the thaw. An
estimated 10 000 lakes, between 10 and 15 per cent of all the lakes on the
Shield, are acidic throughout the year; this number probably rises to 100
000 during the spring thaw. Although these figures are controversial, the
phenomenon of acid discharge is real and is, in many cases, related to pollution
from human activities.

Global warming, another consequence of human activity, will also upset
the freeze-thaw cycle and so change the fate of atmospheric pollutants.
In the next few decades, as the world warms in response to the build-up
of greenhouse gases in the atmosphere, the area of seasonal snow cover will
shrink. What there is will lie for a much shorter time each year. The consequences
will be felt far away. Large changes in the extent of the snow across Eurasia
are linked to changes in the world’s climate, such as the intensity and
timing of the Indian monsoon. According to Tim Barnett, of the Scripps Institution
of Oceanography in California, an increase in the amount of snow over Eurasia
is followed by a poor monsoon.

The absence, or reduction, of seasonal snow will also alter the pattern
in which pollutants are deposited. For example, snow normally protects vegetation
from the dry deposition of sulphur dioxide for several months of the year.
More sulphur dioxide is deposited on vegetation than on snow – perhaps four
or five times as much; trees in full leaf probably receive still more. Without
a protective layer of snow, the vegetation will be subject to a continuous
assault from sulphur dioxide.

Snow reacts less with atmospheric pollutants than does the underlying
land. If the snow disappears, the effect will be twofold: first, in some
regions, pollutants will travel much shorter distances, because dry deposition
takes place much faster on vegetation than on snow, removing these pollutants
from the atmosphere much sooner. Secondly, the surface of the land will
receive pollutants evenly over the year, rather than in one large flush
at snowmelt. No one can predict what the effects will be. It is clear, however,
that northern ecosystems, particularly the coniferous forests, are at some
risk. These forests, which spread over most of the land that is covered
by seasonal snow, are adapted to the influx of nitrogen from the melting
snowpack. Any large-scale change in snow cover is likely to upset the equilibrium
of such an ecosystem.

The importance of snow cover in annual geochemical and nutrient cycles
is unquestioned. What will happen in future decades is a subject for speculation,
but as snow cover shrinks in the years ahead, chemists must remedy their
ignorance of the chemistry of the snowpack.

* * *

Frozen records of atmospheric pollution

MOST researchers investigating the chemistry of the snowpack have concentrated
on snow at high latitudes, sampling not only the snow but also the layers
of ice forming beneath. Because fallen snow in many of these regions does
not melt, but accumulates and evolves into ice, it holds a record of past
atmospheric conditions.

Analysis of ice cores from Greenland and Antarctica provides a record
of changes in the climate and the atmosphere over the past 100 000 years
(‘Frozen assets of the ice cores’, New ÐÓ°ÉÔ­´´, 14 April 1988). In a similar
way, new deposits of snow and shorter cores of firn (the porous, early stage
of ice) and solid ice provide invaluable information on recent trends in
atmospheric chemistry, particularly in relation to atmospheric pollution.
The figure shows the trends in pollution over the past century, which are
visible in Arctic snow. Anthropogenic emissions of sulphur and nitrogen
are probably responsible for these trends. In contrast, the chemical composition
of snow falling in the Antarctic reveals no significant changes as a result
of pollution elsewhere in the world.

The only evidence for human induced changes in deposition comes from
the radioactive contaminants that were locked in the snow after the atmo-
spheric nuclear tests of the 1950s and 1960s. The well-marked horizons defined
by these radioactive isotopes, found everywhere in the Antarctic, provide
an accurate means of dating the recent layers of snow on the continent.

The transformation of porous firn into airtight ice is accompanied by
the trapping of air bubbles. Consequently, there is always a time-lag between
the ages of the enclosed air bubbles and the ice that contains them. This
lag may range from a few years to about a thousand years. Despite this complication,
analysis of carbon dioxide and methane in the bubbles of recently formed
ice are invaluable for reconstructing the extent to which human activities
have altered the concentrations of these gases in the global atmosphere.
Purely anthropogenic trace gases, such as CFCs, are also locked into the
ice bubbles. It may be possible to calculate how much the concentration
of these trace gases has increased in the atmosphere from ice collected
from areas where a great depth of snow has accumulated, and firn changes
into ice rapidly.

Snow and ice also carry a record of volcanic eruptions, which can affect
the climate. When a volcano erupts, it spews out huge amounts of sulphur
dioxide (as much as several million tonnes) as well as dust and other gases.
The sulphur dioxide is subsequently converted to sulphuric acid which is
easy to detect in polar ice and snow.

Polar snow can also provide information on fundamental processes that
take place in clouds. One example is the reaction, well known at mid- latitudes,
of strong atmospheric acids with particles of sea salt in cloud droplets.
The reactions release gaseous hydrogen chloride to the atmosphere. Recent
analysis of polar snow has shown that, during winter, air masses move so
quickly around and across Antarctica that there is not enough time for hydrochloric
acid to rain from the clouds. Therefore the atmosphere at the pole in winter
– and so also the winter snow – does not contain significant concentrations
of hydrochloric acid, although there are clear increases in the summer.

Trevor Davies researches into the chemistry of snow at the School of
Environmental Sciences, University of East Anglia. Robert Delmas, of the
Glaciology and Environmental Geophysics Laboratory, in France, Gerry Jones,
of the National Research Institute in Quebec and Martyn Tranter of the University
of Southampton also contributed to this article.