The calculation of the QMED from POT data is based on the methods outlined in FEH 3 ‘Statistical procedure for flood frequency estimation’ (Section 12). The method is often used as an alternative estimate of QMED when only a short record is available. The inputs to the estimation process are the number of hydrological years of valid data and the POT values. It is also possible to change the dispersion factor. There may be periods for which the data is not valid and these may not always coincide with ‘full’ hydrological years. The valid data and number of hydrological years periods can be identified through the ‘Record Period’, ‘POT Rejected’ and ‘POT Gaps’ which are stored within the Peak Flow Dataset .pt file.
The developers of the POT dataset, the NRFA, confirm that where POT data is provided, it can be assumed that the hydrological year is valid i.e. all POTs within a hydrological year are valid even if there are ‘Gaps’ identified. This assumption, combined with the information about the valid periods can be used to identify the number of valid hydrological years.
Within WINFAP (5.3.9411 and prior) the number of valid hydrological years is calculated using the valid period information based on the number of complete hydrological years present. This may therefore underestimate the number of valid hydrological years used in the calculation. In addition, when 14 or more hydrometric years are identified it will erroneously return the highest POT value. The QMED presented in the ‘QMED Rural Estimation’ form is also the observed value (rather than the rural value).
WINFAP will be updated based on the assumptions above. The revised process will be:
- Identify the largest number of possible hydrological years available based on the ‘Record Period’.
- Subtract for hydrological years that fall within the ‘POT Rejected’ or ‘POT Gaps’ periods.
- Add hydrological years back in where POT data (not rejected) exists during these periods.
All POT data that is not rejected will be included in the analysis.
In rare cases where a small proportion of a hydrological year is identified in ‘POT Gaps’ (for example) and there are no POT in the remainder of the year this logic may underestimate the number of valid hydrological years, providing a conservative (high) value of the QMED. In this case the dataset can be adjusted by editing the ‘Rejected periods of record’ in WINFAP. Note that where values are edited in WINFAP using ‘Exclude within project’, these values will not be included in the calculation, but the number of valid hydrological years will not be updated.