Estimates of peak river discharge are essential for designing and managing hydraulic 9 infrastructure such as dams, bridges, and flood alleviation schemes. Typically these 10 are derived for an assigned annual exceedance probability (e.g. the 1 in 100 year 11 flood), and the provision of accurate estimates is a critical issue in engineering 12 hydrology, affecting both financial cost and human lives. In the UK, practitioners 13 typically apply the Flood Estimation Handbook (FEH) statistical

more »

... od which 14 estimates the design flood as the product of a relatively frequent flow estimate (the 15 index flood, IF) and a dimensionless regional growth factor used for estimating peak 16 flows at higher return periods. For gauged catchments the IF is usually estimated from 17 observations as the median annual maximum flow which has a two year return period. 18 For ungauged catchments it is computed through a multiple linear regression model 19 based on a set of morpho-climatic indices of the basin. 20 While the FEH IF methods provide peak flow estimates that are robust and defensible, 21 they do not readily take into account catchment or rainfall heterogeneity (important for 22 large catchments) or the effect of environmental change on river flows. Successful 23 application to regions outside the UK currently requires a network of good quality, long-24 term flow gauges to underpin the design flood method, not always present in less 25 industrialised regions of the world. 26 With the aim of addressing these limitations, we present and assess a methodology 27 to estimate the IF at national scale using continuous simulation from an area-wide 28 physically-based hydrological model (Grid-to Grid or "G2G"). The new methodology is 29 tested across Great Britain and compares well with estimates of the IF at 550 gauging 30 stations (R 2 =0.91) and similar performance simulating the annual maxima trend over 31 time. The promising results for Great Britain support the aspiration that continuous 32 simulation from large-scale hydrological models, supported by the increasing 33 availability of global weather, climate and hydrological products, could be used to 34 develop robust methods to help engineers estimate design floods in regions with 35 limited gauge data or affected by environmental change. 36 37 42 infrastructure. Commonly, design flows are estimated with statistical models fitted to 43 annual maxima (AMAX) measured at a gauged site (flood frequency analysis). 44 Unfortunately hydrological records are often unavailable at the site of interest or, when 45 available, they are too short to allow reliable statistical analyses. To overcome this 46 limitation a standard approach is to adopt a "regionalization" procedure which 47 64 Bocchiola et al., (2003) provides a summary of some of the most widely used. Broadly, 65 if the site of interest is gauged, the IF can be estimated by direct methods, i.e. from 66 the AMAX time series, using the sample mean (Dalrymple, 1960, Hosking and Wallis, 67 1993, NERC, 1975, the sample median (Robson and Reed, 1999), or using peak over 68 threshold analysis (Chow et al., 1988, Robson and Reed, 1999). If the site of interest 69 is ungauged, a variety of "indirect" methods have been proposed to estimate IF. The 70 most commonly used are empirical methods (Hirsch et al., 1992; Meigh et al., 1997; 71 Kjeldsen and Jones, 2009) that relate the IF evaluated by AMAX measurements to a 72 set of morpho-climatic catchment descriptors such as area, slope, average annual 73 rainfall, land use, etc. These methods include coefficients that are usually estimated develop a robust methodology to help engineers estimate design floods in regions with 324 limited gauge data or affected by environmental change, potentially saving many lives. 353 data in a grid-based hydrological model to estimate spatial variation in changing flood 354 risk across the UK. Journal of Hydrology, 377(3), 335-350. 355 356 Bell, V.A.; Kay, A.L.; Davies, H.N.; Jones, R.G.. 2016 An assessment of the possible 357 impacts of climate change on snow and peak river flows across Britain. Climatic 358 Change, 136 (3).