HR: 0800h
AN: GC11A-0140    [Abstracts]
TI: Regional Probablistic RCM-Estimates of Extreme Rainfall for the UK
AU: * Ekstrom, M
EM: m.c.ekstrom@exeter.ac.uk
AF: School of Geography, Archaeology and Earth Resources, University of Exeter, Amory Building Rennes Drive, Exeter, EX4 4RJ, United Kingdom
AU: Fowler, H
EM: h.j.fowler@ncl.ac.uk
AF: School of Civil Engineering and Geosciences, Newcastle University, Cassie Building, Newcastle upon Tyne, NE1 7RU, United Kingdom
AU: Blenkinsop, S
EM: s.blenkinsop@ncl.ac.uk
AF: School of Civil Engineering and Geosciences, Newcastle University, Cassie Building, Newcastle upon Tyne, NE1 7RU, United Kingdom
AB: Widespread flooding in the summer of 2007 and previously in the autumn/winter 2000/01 caused significant damage to the built and natural environment of the UK, highlighting the vulnerability of UK-infrastructure to extreme rainfall events. There is public concern that these events may increase in the future, possibly due to global warming. Projected rainfall fields from regional climate models are used to better understand how regional rainfall extremes may change in a climate of enhanced greenhouse conditions. However, due to their complex nature, RCMs are associated with a number of sources of uncertainty. By using a large number of RCMs some of this model uncertainty can be quantified. Here, changes to future rainfall extremes for the UK are investigated using 13 RCMs provided by the PRUDENCE experiments. Of particular interest is model performance in the spatial domain, and model results are analysed for each of the nine commonly accepted UK rainfall regions. The PRUDENCE RCMs were nested with 4 different global climate models, although only 2 RCMs were driven by more than one GCM. For each region and model, probability densities are estimated for the 1, 2, 5 and 10-day RCM rainfall amounts associated with the 5, 10 and 25-year return period for a control period (1961-90) and for a future period (following the SRES A2 2071-2100 scenario). Return period magnitudes are estimated using a combination of Regional Frequency Analysis and Extreme Value Analysis, where a Generalized Extreme Value (GEV) distribution is fitted using the method of L-moments to annual maxima series of the rainfall totals. Probability densities are generated using 10 000 samples of return period estimates, derived from boot strap samples of regionally pooled annual maxima series. Probability densities based on RCM rainfall for the control period are validated using observed rainfall, and changes in magnitude between control and future experiments are calculated. This will be presented and results discussed.
DE: 1637 Regional climate change
DE: 1817 Extreme events
SC: Global Environmental Change [GC]
MN: 2007 Fall Meeting