Current warming and likely future impacts - Climate Change: Biological and Human Aspects - page 354

purely hypothetical example in the form of a component of the Earth system, V might
be the annual end-summer volume of ice on some high-latitude island. To continue
this purely invented analogy, as the temperature warms the ice volume on the island
decreases. However, as warming continues we may get more ocean evaporation and
hence precipitation as snow, and so the rate of ice depletion on the island slows. But a
point will come with continued warming when the extra snowfall cannot counteract
the effect of the warming and suddenly all the ice on the low-lying area of the island
melts. Indeed, that this can happen quite fast may not come as a surprise given local
albedo effects: less ice and less sunlight reflected by the landscape and so forth and
so even more warming (see Figure 1.8). Snow and ice on the island's mountains are
all that remains and even these continue to decline with even further warming and
snow lines rise up our hypothetical island's mountains.
There are two important points about critical transitions. First, the transition does
not take place about a specific point, but in a zone. This is why the term 'tipping
point' is not appropriate. There is instead a zone of transition. Second, there is a large,
sudden change in the zone of transition. (This is why some do call it a tipping point,
as it conveys the notion of a large change in one variable for a very small change in
another variable.) It is as if the system has two overlapping states: between the area
encompassed by i-iii/ b - c on Figure 6.11b and the area encompassed by ii-iv/near-
horizontal line level with a . The states of the system within these areas might be
considered analogous to areas of comparative stability around a 'Lorenz attractor',
in the parlance of chaos theory, outside of the transition zone.
Other examples of components of the Earth system thought to potentially exhibit
critical transitions include the Meridional Atlantic Turnover, Greenland ice-cap melt
and the south-east Amazon ecosystem. Other examples where the whole Earth system
went through a critical transition include the onset and termination of the Toarcian
and Eocene carbon isotope excursions.
The question that arises is whether or not it is possible to predict that changes in
a system are driving it towards a critical transition. In 2008 a team of German and
Dutch researchers, led by Vasilis Dakos and Marten Scheffer, examined data from
eight ancient abrupt climate change events. They showed that all were preceded by a
characteristic slowing down of climate fluctuations and that this slowing down started
well before the actual shift across the threshold. By slowing down they mean that far
away from the transition zone and abnormal event (be it a drought or whatever) sees
the system quickly return to normal. However, near the transition zone the system only
returns to past normality slowly (e.g. the drought persists for a while). Such slowing
down, measured as what mathematicians call increased 'autocorrelation', Dakos et
al. showed to be a hallmark of climate critical transitions. The following year this
question was reviewed in Nature with the inclusion of other possible mathematical
diagnostics (Scheffer et al., 2009).
The problem is that a long-term environmental database needs to be analysed
to see this change in autocorrelation and other diagnostic signals. But long-term
research projects are time-consuming and so comparatively expensive. Consequently
considerable foresight, as well as the willingness to invest in the research, is needed
to determine in good time which key long-term monitoring databases need to be
established.