 |
Rainfall extreme value analysis and IDF relationships (Patrick Willems) |
 |
|
Intensity/Duration/Frequency-relationships (IDF relationships) give the relation between the rainfall
intensity, the duration of this rainfall and the frequency with which the combination of the two previously mentioned
variables occur (the frequency is the inverse of the return period).
They can be applied to determine the critical mean rainfall intensity in terms of the duration of this mean intensity;
this is typically done for urban drainage design applications where the duration is taken equal to the concentration time in the drainage system.
To construct the IDF relationships, statistical extreme value analysis is required.
The Peak-Over-Threshold (POT) or Partial-Duration-Series (PDS) is applied, with Q-Q plots to evaluate the tail of the extreme value distribution,
to define the optimal threshold, and to calibrate the Generalized Pareto distribution (GPD) based on regression in the Q-Q plot (QQR method).
Three types of IDF relationships for Flanders have been developed:
- the IDF relationships for Flemish design applications determined in 1996,
based on the rainfall data at Uccle/Ukkel (provided by KMI/IRM) for the period 1967-1993 with a time step of 10 minutes.
- the IDF relationships for the Netherlands based on rainfall data at De Bilt (provided by KNMI) in cooperation
with the Department of Sanitary Engineering of the Delft University of Technology (the Netherlands).
The comparison between the Flemish and Dutch IDF relationships (Uccle versus De Bilt) shows that the two rainfall climates are identical.
Therefore the same design rainfall can be used in both countries.
- composite IDF relationships for summer and winter seasons and two storm types (convective thunderstorms and cyclonic/frontal storms),
based on 10 minute precipitation depths at Uccle/Ukkel for the period 1967-1993.
For every aggregation level in the range from 10 minutes to 15 days,
extreme precipitation depths are described by a two-component exponential extreme value distribution.
The two components of the distribution are associated with two different storm types.
The calibration of the IDF relationships is based on scaling laws for rainfall.
Publications:
Willems P. (2000). Compound intensity/duration/frequency-relationships of extreme precipitation for two seasons and two storm types.
Journal of Hydrology, vol. 233, 189-205
<top>
|
|
|
Rainfall input requirements for hydrological
calculations (Patrick Willems) |
|
|
Rainfall is the most important input for many hydrological and hydraulic design calculations and modelling applications.
Long historical rainfall series can be simulated in the hydrological model,
and the simulation results statistically post-processed in order to derive statistical properties on the hydrological variables of interest.
When such approach is used in combination with detailed models, model calculation times are huge.
Therefore research has been set up in order to define which kind of simplifications can be made with respect to the rainfall input.
In this research, the accuracy of the modelling results and the calculation effort are balanced. The balance point might be different for different applications.
The different types of rainfall simplifications considered are composite design storms, short selected rainfall series and modified single storm events.
In many cases, the best approach is the use of a simplified model in combination with continuous long-term simulations.
Well-calibrated simplified conceptual models can reach almost the same accuracy as the corresponding detailed physically-based models,
but have much smaller calculation times. The simplified models can furthermore be used to select or compose the proper rainfall input for detailed modelling.
For use in urban drainage design,
composite storms were created based on the IDF relationships,
following the principle of the 'Chicago storm'. For each
time interval considered symmetric around the centre of
the storm, the rainfall volume of the storm is equal to
the rainfall volume according to the IDF relationships
(volume = intensity x duration). The name 'composite storm'
originates from design applications where one single storm
can be used to assess the peak discharges and peak water
levels for a given return period for all points (with
different concentration times) along the system. In order
to validate the approach, results of the composite storm
simulations were compared with the results from the simulation
of the full historical rainfall series, for several combined
sewer systems. Based on the composite IDF-relationships
for two seasons and storm types also seasonal composite
storms were created.
Publications:
- PhD Guido Vaes (1999) : The influence of rainfall input and model simplification
on combined sewer systems design.
- Full paper 'Rainfall input requirements for hydrological calculations' (2001) in
Urban
Water.
- Full paper 'Design rainfall for combined sewer system calculations : comparison between Flanders and the Netherlands' (2002) at 9th International Conference on Urban Drainage, Portland, US.
- Full paper 'Design rainfall in the new Flemish design guidelines' (2003) at 6th international workshop on precipitation in urban areas, Pontresina, Switserland
- Full paper ' The use of design rainfall in the new Flemish urban drainage guidelines' (2003)
at the International conference on hydrology : science & practice for the 21st century, London, UK.
Posters:
Download composite storms:
<top>
|
|
|
Historical trends and cycles; trends due to
climate change (Patrick Willems, Victor Ntegeka) |
|
|
In 1999 the digitalisation of old rainfall records with
continuous rainfall registrations since 1898 at the main
station of the Royal Meteorological Institute of Belgium
(KMI/IRM) at Ukkel/Uccle was completed by RMI in a joint
research project for the water authority AMINAL. This
resulted in a unique rainfall series with 10 minutes rainfall
intensities over a period of 100 years (since 1898). Based
on this long-term and high frequency rainfall series, analysis of trends and oscillations (both short-term and multidecadal cycles) have been carried out.
In an on-going research project (CCI-project)
for the Belgian Science Policy authority / Federaal
Wetenschapsbeleid) a more detailed investigation is being
made on the historical trends and cycles in the rainfall
intensities as well as in the evapo(transpi)ration series.
This is done after combining historical rainfall series
with simulation results from climate models. Particular
focus is given on predicting the impact of climate change
on the trends in future rainfall and evapo(transpi)ration
Publications:
- Vaes, G. Willems, P., Berlamont, J. (2002). 100 years of rainfall registration: are there trends?, Water Science and Technology, 45(2), 55-61
- Blanckaert, J., Willems, P. (2006),'Statistical analysis of trends and cycles in long-term historical rainfall series at Uccle', Proceedings of the 7th International Workshop on Precipitation in Urban Areas, St. Moritz, 7-10 December 2006,124-128
- Ntegeka, V., Willems, P.(2008). 'Trends and multidecadal oscillations in rainfall extremes, based on a more than 100 years time series of 10 minutes rainfall intensities at Uccle, Belgium, Water Resources Research, 44, W07402, doi:10.1029/2007WR006471
Website CCI-HYDR project on
'Climate change impact on hydrological extremes along
rivers and urban drainage systems'
<top>
|
|
|
|
