Climate variability and
climate change
The verification of climate change as a real phenomenon, and
the quantification of its progress, is an extremely difficult challenge in
terms of processing information from multiple locations and of widely
different types. So, for example, over a period of several years, new record
summer temperatures may be observed at a range of European locations, along
with an increased prevalence of serious flooding across all seasons, while the
annual arrival times of the first migrant birds, and the first-flowering times
of certain plants, may be observed to drift steadily forward.
While a comprehensive analysis of all available information
represents an unattainable ‘holy grail’ for climatologists, the idea motivates
a range of more focussed problems at different levels, representing steps in
the right direction.
Hence, possible research topics for a PhD programme may be
outlined by more well-defined questions about specific aspects of climate
change. For example:
- increased variability is anticipated in the UK
rainfall climate, producing more extremes in both directions – i.e.
more flooding, and more droughts. An analysis of UK
rainfall data focussing specifically on both types of extremes would be a
novel starting point for a project;
- rainfall data collected over a network of approximately
200 sites in the UK
strongly suggest an upward trend in the magnitude of the most extreme
(large) rainfall events when the sites are considered as providing
independent records. The records are not independent however. In
order to verify/quantify any upward trend in extremes, appropriate models
are needed, incorporating the dependence structure between sites. A
starting point for a project would be to employ and build on existing
multivariate
methods, with the aim of realistically assessing the evidence provided by
the UK
rainfall data;
- the
identification of a key set of variables for the quantification of future
climate change is a valued aim for climatologists, and one with a strong
component of statistical modelling required. A climatologist might ask,
for example, what events would have to be observed over the next ten years
to help assess whether perceived climate variability represents
change rather than chance. Hence, an objective of a PhD
project might be the identification of an ensemble of key target variables
for observation, perhaps drawn from a wide range of possible candidates,
with the aim of constructing
a sensitive and objective tool for quantifying change.
As indicated by the suggested possibilities above, there is
a wide scope for the student to choose and guide the topics of study in this
project. The research would be likely to have an emphasis on Extreme Value
Methodology and Bayesian Inference, and to involve a good deal of computational
work. The aims are applied in nature, but can naturally lead to more
theoretical work if the student is inclined to pursue this.