Hayley J. Fowler & Marie
Ekström talk with ScienceWatch.com and answer
a few questions about this month's Fast Moving Front in the
field of Geosciences.
Article: New estimates of future changes in extreme
rainfall across the UK using regional climate model
integrations. 1. Assessment of control
M;Kilsby, CG;Jones, PD
Journal: J HYDROL, 300 (1-4): 212-233, JAN 10 2005
Addresses: Univ Newcastle Upon Tyne, Water Resource Syst
Res Lab, Sch Civil Engn & Geosci, Cassie Bldg,
Newcastle Upon Tyne NE1 7RU, Tyne & Wear,
Univ Newcastle Upon Tyne, Water Resource Syst Res Lab, Sch
Civil Engn & Geosci, Newcastle Upon Tyne NE1 7RU, Tyne
& Wear, England.
(addresses have been truncated)
Why do you think your paper is highly
We believe that this paper and its companion paper, "New estimates of
future changes in extreme rainfall across the UK using regional climate
model integrations. 2. Future estimates and use in impact studies,"
(Ekström M, et al., Journal of Hydrology, 300[1-4],
234-51, 2005), have been highly cited due to a combination of the
popularity of the topic (i.e., climate extremes) and the fact that the
extreme value and regional frequency analysis methodology used within the
paper is well-described and can easily be repeated by other researchers for
other regions, using different regional climate models.
It is, to our knowledge, one of the first examples of the use of regional
frequency analysis in extreme value analysis to provide more reliable
estimates of high return period rainfall events, particularly using pooling
of regional climate model information. This method allows more accurate
assessment of changes to rare events.
Does it describe a new discovery, methodology, or
synthesis of knowledge?
The paper draws together methodologies from the fields of hydrology and
climatology. It is a pragmatic paper where methodologies commonly applied
within hydrological research (i.e., regional frequency analysis and extreme
value theory) are applied within a climatological framework where, more
frequently, methods such as percentiles have been used to describe changes
Would you summarize the significance of your paper
in layman's terms?
One of the few options available to climate scientists, in terms of
studying the impact of future climate change on rainfall, is to use climate
models. However, climate models are not perfect representations of the real
climate and therefore, despite their complexity, the climate models need to
be evaluated so that we can identify how well they are performing.
To assess the performance of the models, we can compare patterns in space
and time of observed rainfall, with the patterns estimated by the climate
model for current climate conditions. In this paper we have studied how a
particular climate model, HadRM3H, represents extreme rainfall in different
regions of the UK. In our paper we show that the climate model does indeed
capture the expected observed magnitudes of rainfall extremes for large
parts of the UK. This is an important validation of the climate model's
ability to simulate extreme rainfall and a necessary requirement if we want
to use the climate model to investigate how rainfall may change under
future climate conditions which we published in the companion paper
(Ekström M, et al., 2005).
This has been used in interim policy guidance on Flood and Coastal Defence
Appraisal by the UK Government Department for Environment, Food and Rural
to produce new indicative sensitivity ranges for peak rainfall
intensity. The study was also cited by the Intergovernmental Panel on
Climate Change's 4th Assessment Report in 2007.
How did you become involved in this research and
were any particular problems encountered along the way?
This research was conducted under a European project called SWURVE
(Sustainable Water Uncertainty, Risk and Vulnerability in Europe) funded by
the 5th Framework Programme. It is a natural extension of
previous research by Fowler HJ and Kilsby CG, "A regional frequency
analysis of United Kingdom extreme rainfall from 1961 to 2000,"
(International Journal of Climatology, 23, 1313-34, 200) and
Fowler HJ and Kilsby CG, "Implications of changes in seasonal and annual
extreme rainfall," (Geophysical Research Letters, 30, 1720,
"Although climate models do not
provide predictions for the future, they
provide scenarios under which users can test
their infrastructure system (e.g., water
supply, storm drainage network, etc.) and
optimize their robustness and preparedness
for different possible futures under climate
These papers were the first applications of regional frequency analysis to
examine changes in observed extreme rainfall, providing the first evidence
of statistically significant changes to UK extreme rainfall during the
1990s which may be a result of global warming. The natural extension of
this work was to apply the same methodologies to examine whether climate
models were able to simulate realistic extreme rainfall and to estimate the
potential changes in extreme rainfall over the long term for management
under climate change.
It should be noted that climate models are simplifications of the real
climate and therefore models provide a "model climate projection" and not a
real future climate or "prediction." As different climate modelling groups
use different ways of creating their models (in terms of representing the
physics and chemistry—we call these "parameterizations"), the models
will give different results. Because there is no definitive "climate
model," it is difficult to say that one model is better than another.
Therefore we need to consider results from many different models to ensure
that we aren't providing a biased result in terms of representing the
The difficulty lies in representing the uncertainty or the range of results
from different models in a sensible way. Are all models equally likely or
are some better than others and how can we tell? These are large questions
which we have continued to work on in our more recent research.
Where do you see your research leading in the
Climate model intercomparison studies that don't attempt to elucidate the
reasons/causes for inter-model differences are outdated. Future work on
deriving rainfall estimates—extremes or otherwise—need to
encompass a greater understanding of the physical causes. We know that
different models give different results, but now we need to investigate why
this is the case so that we can start to reduce/understand the uncertainty
in the differences between models and thus, our estimates of future
Do you foresee any social or political implications
for your research?
This research has already been used in interim policy guidance on "Flood
and Coastal Defence Appraisal by the UK Government Department for
Environment, Food and Rural Affairs within the UK." The issue of changing
extremes of weather is highly politicized and has huge implications for
social systems through changing community resilience.
Climate extremes happen now and will continue to occur in the future,
perhaps with even greater frequency. Flooding and the sustainable use of
water is not just a problem for future climates; it exists now in many
parts of the world. Any tools that can help organizations or governments to
plan for and adapt to possible future changes to already overstretched
resources are urgently required.
Although climate models do not provide predictions for the future, they
provide scenarios under which users can test their infrastructure system
(e.g., water supply, storm drainage network, etc.) and optimize their
robustness and preparedness for different possible futures under climate
Dr Hayley J. Fowler
Reader in Climate Change Impacts and NERC Research Fellow
School of Civil Engineering and Geosciences
Newcastle upon Tyne, UK Web
Dr Marie Ekström
Research Projects Officer
CSIRO Land and Water
Canberra, Australia Web