Supplementary memorandum submitted by
the Met Office (FL 120a)
1. The Met Office is grateful for the opportunity
to provide oral evidence to the Environment, Food and Rural Affairs
Select Committee inquiry on the summer 2007 floods. At the session
on 12 December, and also through Marek Kubala's letter of 18 December
2007 to Professor Mitchell, you invited witnesses to provide further
written evidence, either to clarify or build on the oral evidence
submitted. The Met Office would like to offer the following in
support of its oral evidence.
Did anyone provide advice on the impact of large
amounts of rainfall falling on already saturated ground?
2. The Met Office always considers the impacts
of the weather when issuing a forecast, so does take into account
previous meteorological events. For example, Met Office broadcast
meteorologists will often refer to the impact of rain falling
on already saturated ground.
3. However, the Met Office does not have
access to detailed hydrological models or maps and so is unable
to add any specific or local detail over and above the general
statement that rain falling on already saturated ground is potentially
a problem. Better integration of hydrological and meteorological
models would provide better information that would show, in advance,
how catchments will react to weather events.
4. At present only limited use is made of
precipitation forecast data in relation to river flooding. As
the resolution of rainfall forecasts improves as planned in the
coming years, better integration with hydrological models and
mapping of river catchments should allow more timely and accurate
warnings of river flooding.
5. As highlighted in earlier evidence, no
organisation currently has responsibility for providing a pluvial
flood warning service. Such a service would also need to integrate
detailed hydrological and meteorological models in order to make
a useful and timely assessment of risk. This capability could
be developed on the back of our high resolution weather forecasting
model to be deployed in 2009. To provide timely warnings would
require a 24/7 operational capability and robust communication
Do we need to re-assess our definition of return
periods to take into account intense rain fall in such short periods
6. There was some discussion at the evidence
session regarding the relationship between return periods and
rainfall intensity. It might be useful to clarify the fact that
return periods do take into account rainfall intensity.
7. Return periods are calculated for a range
of weather events, including rainfall events, which are specified
in terms of duration and quantity (giving an average figure of
intensity over a known period). So, for every rainfall event it
is possible, in theory, to calculate the associated return period
either for the whole event or for a particularly intense part
8. The concept of a 1 in 200 year rainfall
event being squeezed into a shorter time period is therefore misleading
as the more intense event will itself have its own return period.
9. For example, the table below displays
data for two locations from 20 July 2007. The table broadly shows
that short periods of intense rainfall are more common than long
periods of intense rainfall. It is important to note that these
return periods relate to rainfall rather than flooding. As already
understood by the Committee, even a short period of intense rainfall
can lead to flash flooding if it is falling onto already saturated
ground or impermeable surfaces.
|* The fact that return periods are calculated using historical data sets typically stretching back 100 years means that extrapolation to produce return periods beyond 200 years is probably not justified. Our statistical modelling techniques can produce return periods in excess of 200 yearswe have shown these in brackets for completeness but we quote these as >200 years.
10. To illustrate the timing and intensity of the rainfall
on the 20 July, hourly rainfalls for Pershore College (Worcestershire)
and Brize Norton (Oxfordshire) are given below.
11. The models and statistical analysis used to produce
the return periods above work well in a stable climate system.
12. However, we know that climate is changing and therefore
the past is no longer a useful guide for the future, in terms
of weather patterns. In order to translate the impacts of climate
change into weather patterns and return periods, more work is
required. The Met Office Hadley Centre continues to conduct world-leading
research in these areas, but we do already know enough to suggest
that the decisions we make now and in the future must be made
in the context of a changing climate.
13. In the future, average winters are likely to be wetter
than they are now and summers are likely to be drier. However,
the expected intensity of severe weather increases in both seasons,
which will have an impact on our flood defence strategies and