UK Flood estimation Flashcards
What are the different ways of estimating floods described in the Flood Estimation Handbook (FEH)?
The FEH presents two alternative routes to flood estimation: 1) fitting probability distribution to river flow data; and 2) rainfall-runoff modelling driven by statistical estimates of rainfall for “design events”.
Briefly describe the term “index flood”. What is the return period of an index flood as described in the FEH?
It is the median annual maximum flood flow, QMED. This is the flood that is exceeded on average in exactly half of all years, thus it has a return period of 2 years and has an annual exceedance probability of 0.5.
How can you estimate QMED in absence of adequate data?
QMED can be estimated from an empirical formula based on catchment descriptors.
Briefly describe the catchment descriptors used to calculate QMED for an ungauged catchment.
AREA
SAAR - Rainfall averages for UK over the 30-year 1961-1990, interpolated to a 1 km grid.
FARL - Any reservoirs or lakes within a catchment will tend to have some effect on flood response, but it is those directly linked to the stream network that are most likely to produce an attenuation effect.
The Flood Attenuation by Reservoirs and Lakes (FARL) index, developed for the FEH, provides a guide to the degree of flood attenuation attributable to reservoirs and lakes in the catchment above a gauging station.
Values close to unity indicate the absence of attenuation due to lakes and reservoirs whereas index values below 0.8 indicate a substantial influence on flood response.
BFIHOST - Hydrology of Soil Types (HOST) dataset
What are L-moments?
L-moments are summary statistics for probability distributions and data samples.
They are analogous to ordinary moments – they provide measures of location, dispersion, skewness, kurtosis, and other aspects of the shape of probability distributions or data samples – but are computed from linear combinations of the ordered data values (hence the prefix L).
Regular moments include the squares (for variance) and cubes (for skewness) of the data – L-moments only use linear combinations of the data.
What are the different regional features that are used for the pooling of catchments according to the FEH?
AREA, SAAR, FARL, FPEXT (% area of floodplain extent)
Why do we use L-moments?
A single very large flood can greatly change the sample moments, and thus greatly influence the fitted distribution.
This is undesirable because those large floods are also often quite uncertain. We don’t want our estimates overly influenced by one uncertain data point.
L-moments provide measures of location, dispersion, skewness, kurtosis which are robust, not influenced by extreme outliers.
