Regionalised impacts of climate change on flood flows: uncertainty analysis. Milestone report 5. Revised November 2009
Kay, A.L.; Crooks, S.M.; Prudhomme, C.. 2009 Regionalised impacts of climate change on flood flows: uncertainty analysis. Milestone report 5. Revised November 2009. Defra, 71pp. (CEH Project Number: C03037, R&D Milestone Report FD2020/MR5)Before downloading, please read NORA policies.
Restricted to NORA staff only
This milestone report for project FD2020 ‘Regionalised impacts of climate change on flood flows’ describes the analysis undertaken to assess the potential level of uncertainty, due to various assumptions and simplifications necessary to develop the project’s ‘scenario-neutral’ approach to regionalisation. It relies on three previous project milestone reports, describing the hydrological models, the catchments modelled and the model calibration (Crooks et al. 2009), the development of the sensitivity framework (Prudhomme and Reynard 2009), and the identification of flood response types (Prudhomme et al. 2009a). The scenario-neutral approach required that the monthly changes in precipitation and temperature suggested by current Global Climate Models (GCMs) were distilled down into a ‘simple’ sensitivity framework, using single harmonic functions (i.e. annual sine-curves with a single peak and trough). These 4200 ‘scenarios’ (525 precipitation x 8 temperature / potential evaporation) were then applied to baseline catchment time-series using the delta change method of downscaling, and run through the catchment hydrological models. This resulted in the production of flood response patterns, representing the response of each catchment to the prescribed sets of changes in precipitation and temperature / potential evaporation in terms of the percentage change in flood peaks at four return periods. The main aim of the uncertainty analysis is to assess whether values obtained from the flood response patterns will consistently over- or under-estimate the impact of climate change scenarios. The uncertainty analysis thus addresses the following factors: 1. Assumptions made for sensitivity framework development; 2. Use of a fitted harmonic instead of monthly factors; 3. Use of the simple delta change method of downscaling; 4. Natural variability. Due to the number of factors investigated, the analysis is performed on a small subset of catchments, chosen to be as representative as possible of the nine flood response types found in Great Britain (described as ‘Damped-Extreme’, ‘Damped-High’, ‘Damped-Low’, ‘Neutral’, ‘Mixed’, ‘Enhanced-Low’, ‘Enhanced-Medium’, ‘Enhanced-High’ and ‘Sensitive’). There is one catchment modelled with the PDM hydrological model (at a daily time step) for each of the nine flood response types, for which the full uncertainty analysis is performed. In addition, there are four catchments modelled with the CLASSIC hydrological model (at a daily time step), representing four of the flood response types, for which a subset of the analysis is performed. The results show that the level of uncertainty from different factors varies significantly between catchments. For some catchments the overall level of uncertainty varies little with return period, whilst for others it increases / decreases with return period. The four CLASSIC catchments show a similar pattern of uncertainty to that for the corresponding PDM catchments. However, each of the CLASSIC catchments has a higher level of uncertainty than the PDM catchment of the same flood response type. This probably reflects the larger catchment area of the CLASSIC catchments. Generalising the catchment results to their flood response types suggests that ‘Neutral’ catchments will have the lowest level of uncertainty and ‘Sensitive’ catchments will have the highest level of uncertainty. The different levels of uncertainty for the different catchment types are compatible with the underlying climatological and hydrological differences between their flood response types. Despite the small number of catchments investigated here, the fact that the results are physically reasonable, and the similarity of the results for comparable PDM and CLASSIC example catchments, gives confidence in the extension of the results to catchment type. The next step is to develop guidance on what level of uncertainty to allow, according to flood response type and return period. The potential effect of catchment area on the level of uncertainty will also have to be borne-in-mind.
|Item Type:||Publication - Report (UNSPECIFIED)|
|Programmes:||CEH Programmes pre-2009 publications > Water > WA01 Water extremes > WA01.3 Quantification of uncertainties, trends and risk of extremes|
|CEH Sections:||Harding (to July 2011)|
|Funders/Sponsors:||Defra, Environment Agency|
|Additional Information. Not used in RCUK Gateway to Research.:||Joint Defra/EA Flood and Coastal Erosion Risk Management R&D Programme. Project FD2020. This report is available from the Defra website: http://randd.defra.gov.uk/Default.aspx?Menu=Menu&Module=More&Location=None&ProjectID=13958&FromSearch=Y&Publisher=1&SearchText=FD2020&SortString=ProjectCode&SortOrder=Asc&Paging=10#Description|
|Additional Keywords:||Hydrological modelling; uncertainty; climate change impacts; delta change factors; regional climate models; global climate models|
|NORA Subject Terms:||Meteorology and Climatology
|Date made live:||09 Mar 2010 15:08|
Actions (login required)