The importance of environmental statistics

Environmental science has come to the forefront of public debate in recent years with the recognition of the potential effects of climate change by governments. There is an active environmental lobby through organisations such as Greenpeace and Friends of the Earth and a number of high profile issues have raised public interest. Examples are problems with radioactive waste (from Dounreay to the South Pacific), widespread contamination from the failed nuclear power plant at Chernobyl, disposal of the Brent Spar oil platform, oil spills such as those from the Exxon Valdez (Alaska) and the M. V. Braer (Shetland), flooding in the UK (see example from 2000), storm damage in the UK (see an example from 2002), algal blooms on lakes (see an example from Loch Leven in Scotland), siting of wind farms, protection of wilderness areas and alignment of new roads.

To look at a historical example, London was concerned about sulphur pollution in air from about 1300. The problem gradually got worse, particularly through the period of the industrial revolution, until the smogs in the 1950s, which caused many deaths. As a result, legislation was introduced to reduce pollution levels for the protection of public health. However, the new tall stacks on power stations in the UK had an effect of exporting air pollution across northern Europe in the 1970s and 1980s. This increased the environmental effects of what came to be known as 'Acid Rain'. Only as we entered another millennium was that sulphur problem substantially resolved.

These examples illustrate some of the characteristics of environmental science. Time scales may be very long, as with climate change, or very short, as with oil spills. The spatial scale may be relatively localised, as with a flood, or very extensive, as with Chernobyl. An environmental effect may be traced to a single cause or may be the result of complex interactions. Quantification of the processes and the effects may often be an issue, particularly if measurement data are difficult to acquire, and there are always underlying patterns or cycles in the environment that are reflected in the data collected. Mathematical models have become an important tool to represent complex environmental interactions and many published environmental statistics are the output of such models. Further, since many decisions on environmental management are open to political debate, and influence on public opinion has become a major factor, the results of environmental studies are often contested.

Statisticians working in the environmental sciences may be tackling problems in areas such as the following:

•  Climatology - for example, assessing changes in climate patterns
•  Oceanography - for example, assessing the patterns of temperature in ocean currents and their effects on the weather
•  Extreme event risk assessment - for example, looking at the probabilities of floods in an area or of increasing wave heights which may damage offshore structures
•  Fisheries statistics - for example, assessing the population size and the development stage of fish stocks from landings and sparse sample measurements
•  Environmental model assessment - for example, using sensitivity and uncertainty analyses on models to determine the accuracy of predicted future carbon budgets
•  Impact assessment - for example, assessing the effects of a new factory on the local environment
•  Environmental epidemiology - for example, assessing the effects of air pollution on asthma occurrence
•  Ecology  - for example, modelling population changes of upland red deer
•  Compliance issues - for example, framing sampling schemes to ensure that legislation protecting rivers from excessive pollution is observed
•  Risk assessment - for example, assessing the risk of contamination and the likely environmental recovery from a nuclear accident.

Applications overlap with those for a biometrician, medical statistician, government statistician and, increasingly, economist. As in other fields, one of the attractions of working as an applied statistician is the opportunity it gives to apply generic methods to a wide variety of different applications.