Selection of QMED Method

The QMED can be estimated using a number of different methods. When the Pooled & QMED dashboard is first loaded, the default QMED estimation method is selected according to whether suitable at-site data is available in the NRFA Peak Flow Dataset or supplementary folder. Where possible it is advisable to use at-site data.

The options to incorporate at-site data in QMED estimation is selected by default if:

  • the Site of Interest has appropriate AM or POT data (see below for criteria), and
  • the station is 'suitable for QMED'.

You can override the default selections if appropriate. It is recommended that you assess the appropriateness of using the data taking into account the above as well as other pertinent station details.

Where at-site data is not available or not selected, the default method selected is Catchment Descriptors. Incorporation of local data is normally advantageous hence for most ungauged sites this is normally modified using the Donor Adjustment method. It is advised that you select the most appropriate method for the site.

You can also provide a User Defined value, which is assumed to be the 'as-rural' estimate. This could be back calculated for the observed data using the UAF presented in the dashboard if necessary.

The different methods are summarised below.

QMED from Annual Maxima data

A minimum of 14 years of gauged data is recommended for estimating QMED from Annual Maxima data. QMED is estimated as the median of the ranked Annual Maxima flood series.

QMED from Peaks Over Threshold data

The FEH states that QMED can be estimated from Peaks Over Threshold (POT) data for gauged stations containing between 2 and 20 years of data. This option will be ungreyed if the number of years of POT data is less than 20 years and greater than 0. This will be selected as the 'best' option if there is less than 14 years of AMAX data and the number of years of POT data is greater than the number of AMAX.

Further information can be found in the Flood Estimation Handbook Volume 3, Section 12.3.

QMED from Catchment Descriptors

If the Site of Interest is ungauged or has only a very short record of AM or POT data (less than two years) the catchment descriptor (CD) method is the default method selected for estimation of QMED. This method involves the use of an equation established for essentially ‘rural’ catchments which derives QMED from catchment descriptors. The equation is presented below and further information is available in Kjeldsen et al., 2008.

QMED = 8.3062 \times AREA^{0.8510} \times 0.1536^{ \big( { 1000 \over SAAR} \big) } \times FARL^{3.4451} \times 0.0460^{(BFIHOST19)^2}

In many instances this estimate should then be refined using Donor Adjustment.

QMED from Catchment Descriptors with Donor Adjustment

The Catchment Descriptor estimate of QMED can be refined using Donor Adjustment, as detailed in Kjeldsen et. al., 2014.

A table of suitable candidate stations for use as donor stations is presented on the Donor Adjustment tab.

image.png

This contains the ten stations with the lowest geographic distance (using catchment centroids). The stations are selected from those in the Station Manager, that:

  • are deemed as 'suitable for QMED'; and
  • have an URBEXT2000 value that is less than the threshold previously set. By default this is 0.03, categorised as 'essentially rural'.

Both these criteria can be changed under the 'Show' section. You can show 'All Sites' or ‘only sites suitable for QMED’, as well as filter candidate donor stations based on URBEXT2000. In particular it may be useful to change the URBEXT2000 threshold to include urbanised stations that are close in proximity to the Site of Interest. Note you can also modify the default initial list of 10 donor stations under Options>User Options on the 'Donor Options' tab.

for a standard estimate, by default, the 6 closest stations are selected for use in the estimation of the rural QMED. You can select and exclude donor stations by selecting the relevant station in the table.

If 'use small catchments recommendations by default' is selected when initiating a Pooled & QMED Analysis, 1 donor station will be selected by default as recommended by Phase 2 of the small catchments research, Stewart et. al., 2019.

A colour and symbol coding system has been adopted in the table:

  • Your Site of Interest is presented in turquoise.
  • Blue stations are selected for incorporation into the donor adjustment.
  • White stations are excluded from the donor adjustment.
  • Stations from the supplementary data folder are identified by +.
  • Stations edited in the project are identified by *.

Using the indicative Factorial Standard Error (FSE) value you can evaluate how the uncertainty of the estimate changes as the donor stations are selected or excluded. Note that the FSE presented does not include sampling error relating to the AM data. If these records are long then this can be assumed to be minimal. In general, differences to the FSE are likely to be small (<0.01).

The adjustment is applied to the 'as rural' Catchment Descriptor estimate of QMED. The QMED estimate for a donor station is therefore deurbanised by default using the inverse application of the Urbanisation procedures.

For more information on the methodological background please refer to Kjeldsen et al., 2014.

The donor table and relevant settings can be copied to the clipboard as comma-delimited text using the 'copy' icon and pasted into external software.

QMED from Catchment Descriptors and Channel Dimension

If your Site of Interest is ungauged and you have an accurate measurement of Bankfull Channel Width (BCW) it may be appropriate to use the Catchment Descriptors and Channel Dimension method to estimate QMED. The equation was developed in the Environment Agency R&D Project SC130009 (Dixon et al., 2017). The method is most suited to perennial streams with stable, natural banks and should be used with care. The equation was developed from a much smaller dataset than the CD equation and therefore should not be the preferred method in most cases. In addition, the equation is also sensitive to the way in which BCW is measured.

QMED from Catchment Descriptors and Flow Variability

This method allows you to estimate QMED based on non-flood gauged data available for the station. The use of this equation can significantly reduce the uncertainty in the estimation of QMED when compared with the CD method. Furthermore, the uncertainty is not dependent upon an estimate of contributing catchment area. This equation is therefore useful if there are local gauged data that are not suitable for direct use in the FEH methods due to hydrometric limitations at flood flows. This method requires the following three inputs from gauged data which are available from the NRFA:

  • Q5 flow m³/s
  • Q10 flow m³/s
  • Base Flow Index

For further information on the use of this method, please refer to the WINFAP 4: QMED Linking Equation Technical Note which provides full details of the method and guidance on use.

User Defined Value

A value defined by the user outside of the FEH methods based on alternative approaches.