Area-wide river flow modelling for the Thames Estuary 2100 project: Climate change impacts on flood frequency
Kay, A. L.; Bell, V. A.; Lowe, J. A.. 2008 Area-wide river flow modelling for the Thames Estuary 2100 project: Climate change impacts on flood frequency. Wallingford, NERC/Centre for Ecology & Hydrology, 41pp. (CEH Project Number: C03253)Before downloading, please read NORA policies.
Restricted to NORA staff only
Download (2310Kb) | Request a copy
The overall purpose of the Thames Estuary 2100 project is to develop a Flood Risk Management Plan for the Thames Estuary (including London) over the next 100 years; a period long enough that climate change must form an important consideration. Here, the changing fluvial flood risk induced by storm rainfall is the focus of concern. Since the 1960s, daily precipitation in the UK has tended to be more intense in winter and less intense in summer. The impact of changes in rainfall on river flows will depend on both the nature of the rainfall and the physical characteristics of the catchment draining to the river. For fast-responding catchments, such as those in impermeable or high relief areas, the characteristics of the specific rainfall event are critical. Such catchments tend not to have the deep soils and permeable geology that lead to the long-term hydrological “memory” of larger lowland catchments. Catchments across the Thames Basin are typical of these lowland catchments where the longer-term balance between rainfall and evaporation is particularly important. Against this background this report uses a distributed hydrological model, the Grid-to-Grid (G2G) model, to assess future changes in river flows in the Thames Basin. Use of the grid-based methodology in conjunction with an ensemble of high-resolution climate model output has made it possible here, for the first time, to estimate the spatial effects of climate change on peak river flows across the Thames. The G2G hydrological model has been used with a perturbed parameter Regional Climate Model (RCM) ensemble to analyse changes in flood frequency across the Thames basin, across two 30-year time-slices—Current (Oct 1960 - Sep 1990) and Future (Oct 2069 - Sep 2099)—under the A1B emissions scenario. A comparison of observed flood frequency with derived flood frequencies for the Current period suggested that flood frequency is assessed well by the G2G modelling framework using a subset of RCM ensemble members for some catchments, but over-estimates flood frequency in other catchments. However, other factors need to be borne in mind when making this comparison, such as G2G model error, flow gauge error, and the consequences of artificial influences like abstraction (not included in the G2G) on observed flow records. Across the Thames Basin, changes in flood frequency between Current and Future periods have been analysed, and the results for each ensemble member are presented at a range of return periods. There is considerable variation in the results, by ensemble member, by return period and by location, with areas underlain by chalk generally showing lower percentage changes than other regions. The range of results is quite large. However almost all changes are increases, generally averaging around 5-10% in chalk areas and around 30-50% elsewhere for peak flows with up to a 20-year return period. It is important to note that the possible future climates and river flow estimates encompassed by the RCM ensemble have not been weighted according to quality or likelihood. However, 6 of the 17 original ensemble members have been excluded as they led to unrealistic simulations of the strength of storms and consequently low estimates of current rainfall over the Thames region. In view of the large uncertainty range, it is recommended that the mean and range of percentage change in river flows across the 11 RCM ensemble members are used for decision making. A comparison of the modelled changes in flood frequency with an RCM-based estimate of current natural variability showed that, whilst for some rivers (or parts of rivers) there are few changes outside of the range of current natural variability, for other rivers there are more changes outside of the range. The latter locations could be considered as sites where further monitoring/modelling may provide early warning of statistically significant changes in observed flows, due to climate change. The large estimated increase in future peak flows is discussed in a wider historical context. Trend analyses of observed flow records throughout the 20th century have so far detected no apparent long-term trend in UK flood magnitude. Over recent years (the last thirty to forty), apparent increases in winter precipitation appear to have led to an upward trend in peak flows in many UK rivers: however no such recent trend is evident for the lower Thames. Ongoing improvements to both the hydrological and climate models used in the study should lead to greater confidence in estimates of future changes in river flows across the Thames. Uncertainty in the hydrological model estimates can arise from a range of sources including emissions scenario, model structure (for both the climate and hydrological models), and parameterisation (again for both the climate and hydrological models). For catchments considered to be particularly susceptible to increases in future flood risk, additional analysis using a catchment rainfall-runoff model (such as the Probability Distributed Model) adjusted for local conditions is recommended. However several studies have suggested that the greatest uncertainty comes from sources related to the modelling of the future climate, particularly the choice of driving GCM, rather than from emissions or hydrological modelling. The RCM perturbed-parameter ensemble applied in this study represents the first attempt at deriving fine-scale information consistent with a range of large-scale regional changes which result from this global modelling uncertainty, and demonstrates how these large-scale uncertainties translate into uncertainty in future flood risk. More work is required to determine how representative these results are of the implications of the full range of climate modelling uncertainty.
|Item Type:||Report (UNSPECIFIED)|
|Programmes:||CEH Programmes pre-2009 publications > Water > WA01 Water extremes|
|CEH Sections:||Harding (to 31.07.11)|
|Additional Information:||See also NORA Item 17851 http://nora.nerc.ac.uk/17851/ for a journal paper related to this report.|
|NORA Subject Terms:||Meteorology and Climatology
|Date made live:||14 Jan 2009 10:35|
Actions (login required)